LzLabs vs IBM: The latest from court

IBM alleges that LzLabs illegally reverse-engineered its mainframe software to help develop its Software Defined Mainframe. (SDM).

Switzerland's LzLabs says IBM is challenging principles enshrined in law under the Software Directive of 1991 and that it is using "speculative and unparticularized" claims to thwart a competitor.

The Stack is at court.

July 4, 2024: IBM-LzLabs case arguments close with ending up in the air

The 10-week mainframe computing trial is over but verdict is yet to come

The IBM UK v LzLabs hearing has, finally, closed but, just as the Technology and Construction Court in Fetter Lane, London never did sort out how to cool the court room climate, it will take some time to unravel what has been, to use a reference employed more than once during the hearing, a Gordian knot of technological and legal complexity.

For the defendant, Mark Vanhegan KC reiterated a theme that LzLabs output used in its SDM product was not a copy of pre-existing code. The company only needed interface knowledge to create its functional equivalence and allow customers to move mainframe workloads off IBM hardware to more affordable Linux servers. “What goes in … and what comes out” were the dual requirements to understand, he said.

Vanhegan again baulked at arguments that code “fragments” and tools should not be regarded as part of the IBM ICA Program contract that the claimant says the defendant breached. LzLabs’ use of an IBM mainframe, via UK testing subsidiary Winsopia, and its discovery process into formats, data structures and interfaces was purely in pursuit of interoperability and letting the mainframe “talk” to the LzLabs SDM product. To see this otherwise or argue that component code injected via compilers, copybooks, macros, CSECTs and the like was “highly imaginative”, “bizarre” and “speculative”.

The ‘OST’ right to observe, study and test was enshrined by the EU Software Directive and IBM had never complained of third-party products like ColeSoft z/XCD used to debug the innards of mainframe binary code.

IBM’s legal efforts had displayed a “fundamental failure” to address the alleged breach, even given its employment of the “umbrella term” of mainframe software to extend the purview of the ICA. Macros and copybooks are not ICA programs but elements “baked in right at the very beginning of every customer application”, Vanhegan said. This was “entirely artificial and contrived … as is the whole case”.

The claimant had signally failed to show breaches or a systematic attempt to disregard IBM rights and there was no evidence of IBM program code in the SDM.

Picking up his defence team colleague’s theme, Roger Stewart KC took aim at the case being brought against LzLabs executives Thilo Rockmann and Mark Cresswell, as well as principal investor John Moores.

Pointing to cases with relevant precedent such as Said v Butt and OBG v Allan, Stewart said that there could be no individual responsibility in tort and individuals had believed what they were doing was legal, having taken appropriate expert counsel.

LzLabs and its Winsopia subsidiary had made strenuous efforts, conducted relevant changes to its code of conduct and, where there had been errors, these were not part of a systematic process. There had been no “looking over the shoulders of Winsopia employees” and a culture of compliance had held firm.

The defence also argued that much of the alleged activity was time-barred because IBM had acted belatedly, breaking terms of its own contracts. There had been no dishonesty or deliberate concealment. What we were left with, after 33 days across 10 weeks, was “a tiny handful of allegations”.

For IBM, Nicholas Saunders KC continued his theme that custom applications were “an amalgam” that contained IBM code: they were “an assembly of bits” or “a hybrid thing”.

The Stack will report on the verdict when it lands with an in-depth analysis shared via mailshot to our subscribers. Subscribing is free and can be done below.

July 3, 2024: LzLabs pleads for common sense and openness on IBM mainframe trial’s penultimate day

Defence says logic dictates that claimant’s view of customers’ contractual obligations is impossibly narrow

Day 32 of this tense, dense, impactful case played out in front of a full house with about 50 people squeezed into a court perhaps 20m x 15m and still lacking an acceptably functioning air-conditioning system.

This was the penultimate day of this London trial and was an opportunity for the defence to give its encapsulated version of events and the claimant’s objections to the judge, The Honourable Mrs. Justice O’Farrell.

In short, its argument leaned on the view that IBM’s interpretation of the ICA contract it claims LzLabs and subsidiary Winsopia breached is absurdly reductive. If accepted, it claimed that the full fruition of this interpretation would create a ridiculous imbalance against customers and partners of vendors.

The day began with a plea by Roger Stewart KC, who rejected suggestions of dishonesty against LzLabs executives Thilo Rockmann and Mark Cresswell, and its principal investor, the storied technology entrepreneur John Moores. All are individually named in the IBM claim against LzLabs and Winsopia.

 Stewart began by noting that an allegation of dishonesty always “raises the temperature and particularly when dealing with individuals and reputations”. Also, in the case of Rockmann and Cresswell there was “no economic advantage” to be gained. This pursuit was only intended “to put pressure … on very intelligent men”.

Even IBM had got “cold feet” after issuing what he called a “blancmange of allegations with no specificity” that would be pursued in degree depending on what came out in trial proceedings. It was “frankly,  nonsense” and there had been “no deliberate concealment”, Stewart said.

Returning to his theme of the previous evening, Stewart said that IBM witnesses had been unpersuasive, naming IBM UK & Ireland company secretary and head of legal Emma Wright and IBM CICS director of development Steven Wallin. Both had frequently reached for privilege and been unable to answer questions, he suggested. (On the previous day, Stewart had memorably compared their pauses to those given to characters by the late British playwright Harold Pinter.)

Picking up the defence thread, Mark Vanhegan KC stressed that LzLabs only wanted to understand code interfaces. The Swiss company had neither wished nor needed to copy IBM code in developing its SDM (Software Defined Mainframe) product that allows IBM mainframe workloads to run unmodified on commodity servers.

“SDM is a compatibility layer or adapter,” he said, and there was “no need to copy any service [but only to] respond to the same inputs and outputs.”

Neither did LzLabs wish to reverse-engineer ICA programs but only customer applications. “If you reverse-engineer a customer application, you do not reverse-engineer an ICA program,” he said.

Customer applications are not copied from source code but generated on the fly and only rely on IBM code “fragments” such as compilers to initialise and execute, Vanhegan said. Disallowing this and rejecting the ability to observe, study and test would defy common sense and mean the end of the commercial appeal of software agreements.

“Customers should be free to do whatever they like with their customer applications [that are] the output of IBM tools, not the tools themselves.”

By making this a claim for breach of the ICA Program contract, IBM had been “shadow boxing”. If its claim were to be successful then the results would be highly anti-competitive. 

“I can't emphasise enough how extreme consequences [would be],” Vanhegan stressed. He asked the judge to imagine a world where RBS or another UK bank would need to license executable code to its Jersey office by  “going back to IBM and saying ‘Could I have a licence to do that?’” 

Source code in isolation would be useless if it could not be “disentangled”. “Every customer application ever created will include code generated from the tools,” he said.

Following his colleague in the case Roger Stewart’s lead, Vanhegan went on to criticise the quality of testimony provided by IBM expert witnesses. He asked rhetorically whether “given IBM’s huge resources”, it couldn’t locate a modern  mainframe expert and whether the court should consider the possibility it was because an independent “may not have been able to support IBM’s argument”. He also asked the court to compare the IBM expert witnesses with those of LzLabs, Prof. Alastair Donaldson and David Stephens.

Any company must be allowed to develop competitive products fairly, he suggested, citing the ColeSoft z/XDC debugging tool that competes with an IBM debugger product as an example. All that competition reasonably requires is an understanding of specific interface functionality.

Vanhegan said a customer application couldn’t be created without compilers and software derived from other tools owned by a vendor but “Once you’ve baked a cake, you can’t take the eggs and flour out.” 

It would be complete commercial nonsense if a customer signs up to an agreement, develops an app and can’t then use it in any reasonable way. “Nobody would sign up to this,” he argued.

Later, however, when he objected to the claimant’s earlier metaphor of a “black box”, the judge made clear her view. While she was grateful for explanations of technical matters, such analogies tend to break down, she argued. “It’s one of the reasons why I find analogies unhelpful,” she said.

The case has one more day to run with the conclusion of the defendant’s closing arguments the principal acts set for tomorrow.

July 2, 2024: IBM seeks out breaches, but now it’s LzLabs’ turn…

The claimant in the IBM-LzLabs case sought to identify micro and macro infringements in its closing arguments, but the defendant hit back hard

IBM completed its closing arguments on day 31 of the impactful London suit against mainframe workloads aspirant LzLabs, choosing to finish off first with summaries that went deep into software code and then pursuing a far broader perspective of allegations against the defendant's development processes and governance, or lack thereof.

The day began with the micro perspective. IBM barrister Matthew Lavy KC undertook a close reading of code, arguing that the defendant’s argument that it merely sought to understand software interfaces was not backed up by facts.

Expert witness views on the matter were partial and misleading, he claimed, while some actions “had nothing to do with interfaces” and touched IBM’s source code. Lavy also disputed the notion that “Winsopia merely acted as a postbox [a suggestion which] rather downplays the role … it’s really doing a lot more”.

In other areas Lavy addressed, software had not been “properly” scrubbed, that is, cleansed of IBM code. 

“Scrubbing really is an area where you have to be careful with the evidence,” he urged judge Mrs Justice O’Farrell. While there has been no shortage of debate over their value, unscrubbed CSECTs [mainframe software pointers] had offered “the piece of the puzzle” LzLabs needed for development of its SDM competitive product, Lavy claimed.

A more elevated perspective was next provided by Fred Hobson KC for the claimant. Hobson said that the defendant knew the essential terms of IBM’s ICA contract programme that limit actions with IBM software and had knowingly run the risk of breaching them. This is “a case built essentially on conscious disregard”, he added.

“The entire reason [LzLabs UK testing subsidiary] Winsopia existed and operated a mainframe was the sole purpose of LzLabs developing a competitive product [and feeding into that],” Hobson said. “It had no other business and no other use for the mainframe. The environment couldn’t have been more different than a bank or airline. If a bank or an airline [created a] breach, that would be purely incidental to core business objectives.”

LzLabs had been “playing with fire” and the risk of a breach of the ICA “must have been obvious”. Evidence of this was that there was “acute concern shared widely in Winsopia and LzLabs that [IBM] would take steps to terminate the contract if it learned what [Winsopia/LzLabs] was up to … They wanted to avoid the claimant learning about Winsopia’s role.”

Winsopia acted at all times at the direction of LzLabs and those controlling LzLabs, Hobson alleged. The case was unusual, he added, in that “the counterparty had no autonomy [and] was set up for the sole purpose of doing what [it was] told … If Winsopia acted in breach it would follow that those [actions] were committed at the direction of LzLabs.”

Attempts at operational separation between Winsopia and LzLabs had been dismantled over time since 2013, Hobson said.

Separation constraints had downsized from Winsopia and LzLabs staff being not allowed to be in the same city to “developers routinely visiting Winsopia and, indeed, coding while at Winsopia”. There followed a culture where it was perfectly acceptable to be “looking over the shoulders of Winsopia employees”. Separation protections “were bypassed in myriad ways” and rules were “regarded by [founder John Moores, CEO Thilo Rockmann and executive chairman Mark Cresswell] as impeding the development of the SDM [LzLabs product]”.

That approach is “akin to taking a seatbelt out of a car that isn’t roadworthy in the first place”, Hobson suggested. Although controls were ineffective for scrubbing code and disassembly, LzLabs and co. had pressed on regardless: “They were perfectly content to let this state of affairs continue”.

Turning to individuals, Hobson said that rules seemingly “didn’t apply to” Moores, who had been “allowed to act a middle man in this grey area” between Winsopia and LzLabs. “The picture that emerges,” he said, “is one of systematic failure at management level … The risk of systematic breach of the ICA was perfectly obvious.”

On Moores, who had testified that he only participated in decisions at arms’ length, Hobson said it was “fanciful to suggest” he would have proceeded with the SDM concept without being familiar with the key legal framework. He was the planning “architect and exercised overall control … he was no passive shareholder”. Instead, he was a person who attended most board meetings and had spent 10 weeks of the year in Switzerland and “routinely leaned into the detail”. He had been the force who led to rapid changes to the code of conduct governing the relationship between LzLabs and Winsopia. 

Finally, the defendants' claim that it had been dependent on legal advice was weakened by lack of proof of detail and in any case did not mean that they did not act “recklessly”.

Wrapping up, Nicholas Saunders KC took aim at details including the suggestion that IBM must have known about LzLabs’ relationship with Winsopia and could have looked up its status on the UK’s Companies House record. This was a “non sequitur” as there was no reason to suspect a relationship until much later.

Deep into the day, in reply, Roger Stewart KC for LzLabs began in colourful style, promising to separate “wood from trees”. IBM claims had been “extraordinary”, he said, advising Mrs Justice O’Farrell to also consider “commercial consequences”, the belated appearance of IBM legal claims and whether there had been any deliberate or dishonest concealment on behalf of the defendants.

He also criticised IBM witnesses, suggesting that there had been time to make a cup of tea while witnesses considered their comments and asked for special privilege to be granted before giving answers. These had been comparable to a play by Harold Pinter, he suggested, referring to the British dramatist famous for characteristic pauses in speech.

Addressing the judge, he said: “You saw a succession of witnesses called by the client and none of them, with respect … was in the first full flush of youth.” Evidence inconvenient to the claimant had been “tailored” and none of the witnesses would be contenders to be “winners of the ‘best witness’ award or ‘trying to help the court’ award”.

By contrast, LzLabs had provided reams of source code, GIT repositories and tens of thousands of emails versus little by the claimant. Notably, he pinpointed IBM’s Mark Anzani, a key figure in IBM’s reactive strategy to LzLabs plans. Anzani was “an experienced litigant … who makes it his practice not to keep documents”, he said. There had been “an astonishing imbalance” in efforts made by IBM and Lzlabs, he added.

The case continues with a scheduled conclusion to closing remarks set for Thursday.

July 1, 2024: IBM barristers begin closing arguments and bid to unravel legal Gordian knot in LzLabs mainframe case

Battle of Big Blue and its mainframe workloads challenger enters its final week 

This week marks the last lap of the IBM UK v LzLabs case that has navigated the rights and wrongs of ways to develop an alternative to running mainframe workloads on commodity servers. The case is scheduled to end this Thursday, July 4.

The day began with Nicholas Saunders KC for the claimant IBM revisiting many of the complex themes that have made up this trial and addressing aspects of the 1,000-page-plus volume its rival has assembled for closing arguments.

Saunders began by questioning the accuracy and detail of that text and urging judge, The Honourable Mrs. Justice O’Farrell, to take note of these potential concerns. After one particularly involved critique, she drew laughter when she reminded the barrister in arch fashion: “Of course, I did have the benefit of hearing the case.” Later, she noted: “It’s not surprising that the clients have a different interpretation of the evidence.”

Saunders went on to suggest that the defence had prepared its expert witnesses on likely lines of questioning in such a way that focus was skewed in LzLabs’ favour. This had resulted in speeches being made rather than direct answers to questions given, and had coloured their evidence, leading to “slightly surreal” exchanges.

“I tend to ignore speeches because they don’t particularly help,” the judge assured him but she added the caveat that she was sympathetic to witnesses that were asked to parse “huge” troves of information “under enormous time pressure”.

Saunders also pointed to the defence’s criticism of IBM expert witnesses and potential partiality, noting that one, Michael Swanson, had told them that he had “very very few friends” remaining at IBM. Witnesses for IBM had been “scrupulously” independent and objective, Saunders said. By contrast, some evidence for the defence had been “misquoted” and was “speculative” with, for example, more subtle opinions ventured in person than they appeared in submitted reports.

Missing text messages and the disappearance of LzLabs backer John Moores’ Gmail account were also drawn attention to, and they had been written at a critical juncture in the development of LzLabs’ SDM product, Saunders said.

This disappearance appeared unusual for a witness so experienced in litigation, he suggested. Moores, a veteran tech entrepreneur, had previously lost another high-profile case with IBM over another attempt to penetrate IBM’s mainframe citadel dominance at an earlier company, Neon Enterprise Software. There was no "innocent explanation" for this, it was suggested.

 The nub of this case is whether LzLabs broke terms of IBM’s ICA agreement when the former’s UK test subsidiary, Winsopia, acquired an IBM mainframe. IBM suggests that by digging deep into code and then sharing findings with its parent, Winsopia breached contractual terms. LzLabs has argued that it only examined IBM code for purposes of integration and compatibility in line with the EU Software Directive.

Saunders said this defence is “nonsensical”. He added: “You can’t take the benefits of the licence if you don’t respect the burden [of usage terms].” 

Also, while LzLabs claims that it only amended software in ways that were allowable, Saunders said it has expressly set out to create a new product and its behaviour in creating the SDM product was “very rare in the field”.

With reference to Greek myth, Saunders compared this to cutting the “Gordian knot” and later essayed another comparison, this time of baking a cake by using multiple ingredients. LzLabs had effectively used reverse engineering and debugging in a manner “worlds away” from normal usage, he argued: “People don’t usually disassemble third-party software … This is not the Wild West … It’s a professional business.”

What LzLabs had done was to look inside the “binary black box” of IBM technology and transfer learnings outside the enterprise, Saunders alleged. Moreover, LzLabs had used sophistry in its interpretation of contractual terms to explain its actions.

“What my learned friends are doing is turning what is straightforward, clear language into complex arguments,” he said.

In the afternoon session, in front of a full courtroom of perhaps 50 people listening intently to the finale of this case, Saunders turned his attention to the Software Directive. He argued that LzLabs had gone well beyond narrow exemptions to rules and reasonable means of creating an alternative to IBM’s mainframe hegemony. IBM had a legitimate right to keep its technology non-public, he said.

The case continues and the expected conclusion on US Independence Day and the UK General Election should lead to a notable conclusion in every sense.

June 19, 2024: In IBM-LzLabs case, the heat is finally off… for now

Expert witness stage of mainframe-centric trial completes and now there will be a hiatus before closing arguments

The lawyers, clerks, witnesses, claimants and defendants that have thronged the Technology and Construction Court in London’s Rolls Buildings could, quite literally, finally breathe more easily at lunchtime on day 29 of the IBM-LzLabs trial.

After cross-examination of the final witness concluded just after 1pm, the court emptied and there was respite from the overheated space and suboptimal fan cooling that has provided a toasty and humming sensory accompaniment to charged legal exchanges over the last month and a half. Court regulars will return a week on Monday, 1 July, when The Honourable Mrs. Justice O’Farrell will hear closing arguments for a few days before delivering a verdict in this case that marks a crossroads in the long history of mainframe computing. 

To recap, IBM is the mainframe market’s dominant player but it is arguably threatened by the emergence of LzLabs. The Swiss-headquartered firm, generously bankrolled by the American serial technology entrepreneur John Moores, created a product called SDM that lets mainframe workloads run on commodity servers, thus potentially saving on hardware, software and maintenance costs.

In this case, IBM UK suggests that LzLabs, its testing subsidiary Winsopia and associated executives broke mainframe licensing terms by reverse engineering one of their mainframes, while the defence has claimed that all development work it undertook fell within terms of the EU Software Directive that allows for open interoperability between systems. 

The day followed the pattern set early in the cross-examination of expert witnesses with forensic examination of what LzLabs had done with code, who that code belonged to, and whether the Swiss company had gained an unfair advantage by taking information from Winsopia, its UK testing subsidiary and IBM mainframe customer. 

Matthew Lavy KC for claimant IBM queried David Stephens, an LzLabs expert witness and quickly moved on to a persistent theme of the case: the suggestion that LzLabs had attempted to replicate or mimic the effects of core IBM software services. 

This time the focus fell on examining IMS (the veteran database and information management system underlying transaction processing systems) and CSECTs (the obscure code that acts as signposts or pointers to callable services around which a great deal of attention has focused from the IBM side).

Lavy put it to Stephens that LzLabs was seeking not so much the CSECTs themselves but information contained therein, in order to develop SDM. Stephens concurred.

Turning to IBM’s JCL scripting language, Stephens said he believed LzLabs had gained a procedure from a customer and assumed development had taken place on top of that. Returning to a familiar refrain, Lavy suggested that Stephens was not here applying technical nous but relying on materials such as emails, messages, comments and content derived from the DR (Discovery Request) development system. Stephens agreed.    

Debate moved on to whether JCL could legitimately be modified as Lavy attempted to draw a distinction between a standard user activity and “wholesale” changes that LzLabs allegedly made. However, Stephens pointed to a further nuance: that an independent software vendor might make significant changes to make its programs work optimally.

Finally, we returned to another leitmotif of the case: “scrubbing”, or the extent to which LzLabs thoroughly cleaned CSECTs. Here again, Lavy suggested Stephens’ evidence was gated by the fact that he only had examined emails and the DR. 

“You haven’t looked for evidence but you haven’t seen it,” he stated. But Stephens insisted it was “obvious” that examples shown had come from a customer and not from Winsopia.

The hearing ended shortly after 1pm with a final flurry of suggestions by Lavy for the claimant over signature CSETs. Playing a straight bat to the end of the innings, Stephens said that what was being discussed was purely data and “data has no functionality”.

The case continues on 1 July for closing arguments and is expected to complete at the end of that week, by which time we will know who has succeeded in generating more light than heat.

June 18, 2024: Final expert probed on LzLabs’ mainframe dev actions as IBM trial nears close

‘What On Earth Is A Mainframe?’ author an appropriate last witness for trial that has scrutinised rights and wrongs of Big Iron software coding

On day 28 of the IBM-LzLabs court case in London, the final witness, mainframe expert and author David Stephens, took the stand again. Stephens once wrote a book called What On Earth Is A Mainframe? and, by the end of what is likely to have been the penultimate day of questioning, he had done his best to explain the complex workings of the veteran host computing software environment.

This was another dense day of examining the complex vagaries of software development and associated legal implications but it began with relatively light fare as Stephens, called as an expert witness by defendant LzLabs, went through debugging processes under the questioning of claimant IBM UK’s barrister Matthew Lavy KC.

In one example, Stephens defended his view that there had been no intention to debug IBM code, only a customer program. This recurring theme is at the heart of the case that focuses on whether LzLabs attempted to hitch a “free ride” on IBM’s still lucrative mainframe business by developing its SDM product, or whether it admissibly studied interfaces to create a competitive offering.

SDM is an effective threat to IBM in that it provides the chance for current IBM customers to run mainframe workloads on low-cost servers and a generally more modern environment.  

Next, the court studied IBM CICS mainframe middleware and areas of what Lavy considered a discrepancy between the views of expert witness Michael Swanson for the claimant and Stephens, who found that no IBM program or object code had been sent from Winsopia to its parent LzLabs. Once again, we were in the disputed area of whether development activities included illegal intrusions into valuable IBM intellectual property or simply LzLabs gaining a better understanding of interfaces through honest means. The latter is protected under the EU Software Directive.

Stephens was asked about the use of tools that Lavy said are primarily for resolving networking issues, namely SLIP traps and GTF tracing. Stephens said that in using them, LzLabs developer Brad Taylor had merely tried to understand data structures; the defence claims that IBM regularly encourages the use of such tools among its customers. LzLabs needed to interoperate with CICS and, therefore, developers needed to inspect and “play” with CICS protocols, Stephens said. Lavy said Stephens could not know this and Stephens conceded the point but said it was plausible.

Reverting back to Swanson’s expert testimony and report, he had said that the use of SLIP traps for debugging was a last resort, Lavy noted. However, Stephens said using SLIP traps was not remarkable and “as a mainframer, there is no reason I can’t use them”.

Lavy countered by quoting Stephens’ own articles cautioning on the use of SLIP traps because of their high CPU usage demands. Stephens agreed that high compute needs was a factor in whether the tool should be deployed but he accepted Lavy’s point that very often they are used to diagnose network issues.    

LzLabs and Winsopia could have gone another way to developing the SDM, Stephens agreed, in theory avoiding many of the issues scrutinised in this trial by using conventional mainframe rehosting approaches.

However, he added that this approach could have caused issues where customers had lost source code, for example.

Lavy suggested that owners of mission-critical apps must retain source code but Stephens said that was not the case “and it drives me crazy”. When Lavy suggested there are services for retrieving lost source code, Stephens suggested that they exist but they present challenges and efforts are not always successful.

The afternoon saw a return to the vexed questions of scrubbing, the process by which code was cleansed by removing IBM stub pointers. Stephens felt confident that CPX, a tool created by Winsopia to remove stub pointers, had been an effective tool for scrubbing. Lavy suggested that Stephens was in no better position to comment than factual witnesses. Stephens responded that he was in a better position because he was not being required to recall actions from a decade earlier and had liberty to inspect emails and explore the DR [Discovery Request development process] with a fresh perspective. 

Lavy suggested there had been a “failure” to scrub certain entries but Stephens argued that “failure is a big word”. Winsopia had created the CPX tool with specific design goals, even if it was plausible that it might have been better, “with 20/20 hindsight” to have expanded its capabilities. “I don’t think omitting [that capability] was a failure,” he added.

The next area for jousting was the use of IBM macros, which Lavy said remained the property of IBM. However, Stephens said that the person using the macro to create the load module is effectively the creator. “As with any IBM macro, you don’t change macro source, you just call it,” he said.

Lavy persisted, suggesting that the user is merely changing values in an IBM program but Stephens disagreed. So, effectively, a macro is similar in this sense to an API, he argued, thus bringing the end of the day. It also brought the day back to a central question of the case as to whether LzLabs/Winsopia could justifiably access IBM code, where and how findings were shared between Winsopia, which owned an IBM mainframe, and its owner LzLabs. “I have not seen any evidence of Winsopia sending IBM material out to LzLabs,” Stephens said.

The case continues. 

June 17, 2024: IBM-Lzlabs mainframe case approaches its endgame - future of mainframe migrations in the balance?

Final witness discusses purposes of code disassembly as London trial enters its denouement

The critical IBM UK vs. LzLabs trial is nearing its end and, on day 27 of the tense, high-profile hearing in a London court, the final expert witness continued to testify.

For the defendant, mainframe expert David Stephens was quickly drawn into arguments on the rights and wrongs of code disassembly and product development. From a wider perspective, the LzLabs sees competitive rights to run mainframe workloads at stake.

As background to the day’s exchanges, mainframe engineer Kevin Lynch worked for LzLabs’ testing subsidiary Winsopia and disassembled IBM’s IGZCUST code.

The defence insisted that IGZCUST is an insignificant aspect of a runtime environment module but it is an outstanding example of LzLabs disassembly beyond customer applications. 

Lynch said that he disassembled about 2% of the module in order to investigate a customer program issue and LzLabs’ defence argued that code did not end up in its SDM program and would not in any case have been useful to developers of that product.

LzLabs' SDM (for Software Defined Mainframe) lets customers run IBM mainframe workloads on commodity servers, thus providing a chance to avoid hardware, software licensing and maintenance costs associated with the mainframe world. (Its partners describe it as "eliminating the need to find, modify, and recompile source code of legacy applications. This approach enables z/Architecture binary executable programs to operate at native speeds on x86_64 architecture computers...")

Stephens gave his expert view that disassembly is very common as a procedure to detect and fix errors but Matthew Lavy KC for IBM said the difference was that, in this case, code was being analysed rather than merely disassembled.

There was some disagreement during the hearing over whether Lynch used a tool or hand-coded the IGZCUST disassembly process.

This was important as context to whether Lynch would know whether he was purposely disassembling IBM code. Lavy suggested that if Lynch had used a tool (a scenario he considered more likely) then he would have disassembled everything before finding problematic instructions. Stephens disagreed with this. 

Lavy next suggested that disassembly by hand would be an “eccentric” approach. Stephens agreed but added the caveat that many mainframe people of his generation had a tendency to eccentricity, prompting laughter in court. Lavy suggested that disassembly by hand was possible but not normal, to which Stephens replied that he avoided the manual approach “at all cost”.

Next, attention turned to compiler listings as a way to debug programs. Lavy asserted that LzLabs and associates used IBM listings to develop SDM. He added that, unlike the majority of cases, LzLabs was not using compiler listings to discover what was and what wasn’t working in a program and why, but instead in order to gain insight into mainframe workings.

Stephens agreed with this latter point.

The case continues.

June 13, 2024: Trial enters endgame as latter’s witness plays a strong defence

Professor acting as expert witness for LzLabs rebuffs suggestions of code development wrongdoing

The endgame of the IBM vs LzLabs case is being played out as the penultimate and final expert witnesses respectively took the stand in the Rolls Buildings in London on day 26 of this pivotal trial covering rights to develop around mainframe computing workloads.

The day began with feisty exchanges between Nicholas Saunders KC for IBM and LzLabs expert witness, Professor Alastair Donaldson. 

Prof. Donaldson, an expert in software development and an academic at Imperial College London, offered a robust and forensic defence to repeated suggestions that he was speculating on issues relating to development of LzLabs’ SDM product.

SDM offers a technical off-ramp for users of IBM mainframe applications, allowing them to run existing workloads on commodity x86 servers. 

Prof. Donaldson also issued rebuttals to suggestions from Saunders that he had not closely examined files relating to the case. At one point, Saunders appeared to become exasperated, suggesting that the witness needed to be clear on what he had and had not examined. “I suggest that answer does you no credit”, Saunders said.

But Prof. Donaldson was robust and frequently corrected Saunders on points of technical detail. He also stressed vehemently that he had made best endeavours to understand the case and its context both before and during the trial. 

Pressed on what he had learned from evidence submitted in the case, Prof. Donaldson stated that proceedings had added context but only reinforced his view that LzLabs development was consistent with standard practice in developing an alternative to the IBM mainframe as a vehicle for workloads.

Prof. Donaldson regularly rebuffed IBM counsel, noting that CSECTs and stubs were merely software fragments that did not represent valuable IBM intellectual property.

There was a testy exchange between plaintiff and defence lawyers when Mark Vanhegan KC for the defence suggested that Prof. Donaldson had already amply answered a question.

“You’re suggesting he is wrong?” Vanhegan continued, in an aside to his opposing lawyer. 

“I’m suggesting he doesn’t know,” Saunders replied.

Judge Mrs. Justice O’Farrell intervened to say, “Let’s move on”. (It was also a notable theme of the day that the Judge asked Prof. Donaldson several times for guidance and clarification of technical points.)

Saunders on more than one occasion instructed Prof. Donaldson not to speculate but when the IBM lawyer asked him whether an analysis of a scenario was plausible, Donaldson replied, “You’re inviting me to speculate?”, prompting more laughter.

The edgy feel of the day continued when Saunders suggested that Donaldson had been remiss in not examining key evidential aspects. Donaldson said he had extracted plentiful time from his university day job and had considered context beyond the strict terms of the case… including listening to a podcast by IBM product management executive Roland Koo.

Evidence throughout had been “illuminating”, he said, but had only served to add nuance and context to his original evaluation.

The final expert witness called was David Stephens for Lzlabs, a mainframe veteran and author of a primer text on the technology at the heart of this case, What On Earth is A Mainframe? 

Cross-examined by IBM, Stephens said he hesitated to define “disassembly” because it is a “loaded word” in this litigation. IBM counsel replied that the witness’s legal opinion was unimportant, to more laughter in court, helping to leaven another day dense with technical minutiae.

Pushed on development techniques, “disassembly is not inherently evil,” Stephens said. He added that it is “something that occurs for fault diagnosis from time to time but not something you regularly see”. He was “cautious” of disassembling code for intellectual property reasons, he added.

The case continues.

June 12, 2024: IBM lawyers peck away but LzLabs expert witness defends views of mainframe workload alternative development 

Prof. Alastair Donaldson insists there was no evidence that IBM code was replicated in LzLabs’ SDM product 

Day 25 of the high-stakes IBM UK vs. LzLabs mainframe rivalry case saw yet more pounding away at an expert witnesses as IBM attempted to unpick evidence supplied by Professor Alastair Donaldson for the defendant. Donaldson, a Professor of Computer Science at Imperial College London, held firm throughout in his views that LzLabs and its Winsopia subsidiary had only made sensible code changes without copying IBM code.

A practical example came early in exchanges when Nicholas Saunders KC for IBM suggested that an email regarding disassembled IBM code sent to John Moores, who bankrolled LzLabs, and Thilo Rockmann, CEO of the company, led to a change in the underlying code for LzLabs’ SDM product. (SDM, or Software Defined Mainframe, is LzLabs’ product which allows customers to run mainframe workloads on commodity servers, thus providing IBM customers with an offramp from the costs, maintenance complexities and software licensing issues relating to being an IBM mainframe customer.) 

The change that was made on the SDM was not “remarkable” after that email exchange, according to Prof. Donaldson’s expert report. He added in court that he does not believe that the sending of the module by Winsopia engineer Kevin Lynch to LzLabs was important to the change in the SDM because it was “so obvious”.

Saunders, reverting to a theme that has been common in witness cross-examination in this case, said Donaldson was “speculating” and stated that “Mr. Lynch disassembled that code and provided comments in his email on aspects of it … he provided a guide as to how it was being run.”

Saunders added: “Where you say in your report that ‘due to the straightforward nature [of the change], I do not regard it as remarkable’, you are putting your own perspective on that, and that is not a reliable guide to what is actually happening.”

Donaldson replied: “These [facts] are not related … It is true [that some code] was rewritten in C after this email but I have not seen evidence that disassembly [of IBM code] could have been useful … I am highly confident that this was not useful.” 

Saunders tried to push Prof. Donaldson to admit that he cannot know or be confident that this email did not trigger the change, nor that the information was useful in making the change. Prof. Donaldson stuck to his position that the change made was obvious to any decent programmer, especially one such as Roger Bowler, the experienced mainframe expert who made the change.

“I have no idea who was shown what and who discussed what with who but, based on my experience, I don’t see any way that assembly code could have informed implementation of the C placement,” Prof. Donaldson said. 

“I do not think it’s credible that this was useful to Roger Bowler, who, I expect, is a highly experienced C programmer and would have had no trouble implementing this [change]. It doesn’t seem credible that looking at assembler code would be useful for implementing this simple function.”

Later, Prof. Donaldson defended his position and stressed that “generally, I tried not to get into details of what LzLabs should or should not do [and instead] tried to look at what LzLabs did and if they were solving legitimate technical problems”. 

Continuing to probe, Saunders showed the witness comparisons between SDM code and tables produced by Winsopia based on IBM code. 

“You say in your report that the SDM code does not reproduce IBM code. I suggest that is wrong,” he said. “You see it here.”

Donaldson replied: “What I see here is SDM code has clearly been informed by the information in the table and that has arisen from study of IBM code in whatever form, but I don’t see reproduction of IBM code in SDM.”

Questioning next turned to the “airlock” network access protection to the Winsopia mainframe. Prof. Donaldson agreed that “it isn’t possible to know if some backdoor existed” that would allow access to a “bad actor”. 

But, he added, it was “clear from evidence that this installation was taken seriously and there are lots of audit trails”. 

The case continues with Prof. Donaldson, the penultimate expert witness, due to give more evidence tomorrow.

June 11, 2024: Mainframe rivals LzLabs and IBM grind away at credibility of respective expert witnesses

Both sides suggest that opposing experts are not modern mainframe specialists 

LzLabs continued its attempts to chip away at the credibility of expert witnesses representing IBM on day 24 of the trial between the warring mainframe rivals. Later, IBM took its chance to question the expertise of LzLabs’ first witness.

The day began with the continuation of an attritional grilling of academic Jon Weissman, a Professor of Computer Science at the University of Minnesota. Early questioning by Mark Vanhegan KC for the defendant focused on the LMD (Load Module Decompiler), a tool developed by LzLabs testing subsidiary Winsopia as a means to migrate customer applications to an x86 commodity server. By moving to an x86 environment, customers can save on mainframe hardware, software and maintenance costs versus being tied to an IBM mainframe and ecosystem.  

The LMD had a two-stage filter for C-SECTs, that is, software pointers for callable services. One was a execution list and the other a compiler test. Weissman only identified the former in his report to the court and Vanhegan asked why Weissman didn’t mention the other filter in his report. Weissman insisted that this was not because he didn’t understand it, as Vanhegan suggested, but because its role was not clear to him. 

Vanhegan said, “You saw it in the source code and decided not to mention it.”

Weissman replied, “When I see lots of things and if their roles [are] unclear, they may or may not be mentioned.”

Additionally, Vanhegan tried to show an instance of Weissman not being “independent and impartial”, thus reviving a theme of LzLabs’ cross-examination of IBM expert witnesses.

Vanhegan said: “Here you are slipping into arguing IBM case. As a fair and independent expert, you should have said, ‘I have found no evidence that any IBM CSECTs were processed by LMD.’”

Weissman: “That is true and I’m phrasing it a different way. There is a limitation in the process, so there is still the possibility IBM CSECTs were processed.”

 Vanhegan: “There is positive evidence that IBM-supplied CSECTs would not have floated through LMD. You and your support team had over 4,000 examples of LMD output, and you had not identified any one of those as containing a disassembled IBM CSECT.” 

Weissman replied that the comparison process was “extremely difficult and time consuming”. However, he agreed that he and his team had not found “a single example” of LMD containing a CSECT. 

Vanhegan later turned to suggestions by defence expert Professor Alastair Donaldson that he did not find any logs that showed LzLabs employees connecting to the Winsopia mainframe. Weissman agreed but argued that there could be “backdoors”. He conceded however that he had no evidence to show these existed. 

Vanhegan: “Wouldn’t it be fair of an independent expert to say that none of the logs you looked at support any proposition the mainframe was accessed?”

Weissman: “You are probably correct.” 

Finally, after several days of dense cross-examination of IBM expert witnesses, Prof. Donaldson of Imperial College, London took the stand to be cross-examined by Nicholas Saunders for IBM.

It was put to Prof. Donaldson, an expert in software development and testing, that he has no mainframe experience as his experience mostly relates to x86 operating systems. 

Saunders: “When you were first approached, you explained that you didn’t know anything [about] mainframes.”

Donaldson: “I did say that.”

Saunders: “So, where you talk about IBM software, you learned about that in the process of this case. How did you do that?”

Donaldson: “I did that through reading, learning online. I read a helpful book by [fellow Lzlabs expert witness] David Stephens [and the expert witness statement].”

Donaldson agreed with Saunders that he is not someone to speak on mainframe industry “norms and conventions”.

For a section of Donaldson’s report, many documents were provided to him by the law firm Clifford Chance after he asked for them. Saunders suggested to Donaldson that there could be breaches he did not find because he did not ask for certain documents. Also, he suggested that Clifford Chance may not have provided the full extent of the documents.

Saunders: “That’s a significant limitation to this section of your expert report.”

Donaldson: “I don’t agree … Lawyers would be better at going through documents when things are not technical in nature ... I would expect that the lawyers would do just a good of job as me for that particular document task.”

The case continues.

June 10, 2024: LzLabs probes for holes in IBM mainframe expert witness

Defendant casts doubt on knowledge of academic Jon Weissman

Ever since expert witnesses for IBM UK took the stand in the Rolls Building court in London last week, lawyers for the defendant have attempted to poke holes in their credibility. Day 23 of the case was no exception to that trend.

Professor Jon Weissman of the University of Minnesota was attacked from the off with suggestions that he was not an appropriate person due to (a) his lack of mainframe knowledge, and (b) that he is a serial expert witness. Also, the Defence tried to show that he had made several mistakes in his report.

The Defence showed that Weissman had not used a mainframe since he was a teenager, decades before the current z/OS operating system.  However, Weissman countered by saying that he teaches 300 students every year on the evolution of the mainframe in the context of operating systems and software. He has been an expert witness in 51 cases in the US and that has been the “bulk” of his employment in recent years. Other cases in which he was an expert witness are “a long, long way from mainframes”, he conceded.

For the defence, Mark Vanhegan KC suggested Weissman has “acted as an expert in more than 30 cases since 2017”.  

“I haven’t counted,” he replied. 

Weissman said he is “active in publishing and giving talks, workshops and conferences … so, the majority of the work is my academic spere and expert consulting pays well, academic does not … I consider myself a first and foremost an academic; I am active in consultancy, that is true.”

Vanhegan suggested that Weissman didn’t really understand what LzLabs’ SDM alternative to running mainframe workloads does, and that he could not know what was IBM proprietary material. 

Next, Vanhegan showed that Weissmann only spent two days at law firm Clifford Chance’s offices examining the confidential SDM code. Weissman said he didn’t dispute the time logs but noted that there was also a repository where he could get some parts of the code to analyse. Occasionally, he was given a pass with someone else’s name on it, so he believed his research had taken more than two days.

Vanhegan persisted, suggesting Weissmann made errors in his report and was specifically misleading the court when discussing copybook allegations. 

The specific allegation revolved around a statement that Weissman made in his report that Winsopia had given IBM-proprietary information to LzLabs through a ‘Discover Request” that would have been in breach of the use of copybooks. However, Vanhegen showed Weissman that actually LzLabs already had the data before submitting the “Discover Request”, and that the copybook in question was a sample program provided and ‘encouraged’ by IBM to be used by its customers free of charge.  Vanhegan’s asserted that since the information was publicly provided and the timeline incorrect, this meant that it could not have been a breach even though Weissman stated that it was in his report

Vanhegan’s suggested that Weissman did not actually write this section of the report, but Weissman responded that he just didn’t check it properly and reaffirmed that he was responsible for the report submitted. 

“This is my report,” he said. “Everything in here is my mistake and my ownership … I did not mislead the court. Creating the report was complex and some things were missed. It’s not perfect. I tried to acknowledge new information [when it appeared].”  

The case continues.

June 6, 2024: LzLabs defence whips up a crescendo of criticism as epic cross-examination of IBM witness closes

IBM expert witness Michael Swanson was finally able to relax as his four-day cross-examination drew to a close on day 22 of the IBM UK v LzLabs trial. His evidence as to the defendant’s conduct in developing its Software Defined Mainframe (SDM) product will be central the case, due to close in three weeks.

In its skeleton summary, LzLabs had suggested that Swanson appeared “to have no real-world mainframe knowledge” since 1999 and defence lawyers attempted to show that knowledge gap.

Swanson had described LzLabs’ use of the Colesoft z/XDC tools as “invasive” in his report but, in court, lawyers demonstrated evidence of it being widely used by “big players” in the mainframe arena “for the purpose of developing commercial software”. Swanson agreed that using z/XDC for testing and debugging was not an “unusual or uncommon” use of the tool. 

Swanson’s report had frequent “hypothetical” alternatives to what LzLabs said it had done when developing the SDM and, Vanhegan suggested, this sort of conditional thinking was not suitable for a truly independent expert.

“You make the assertion there must have been invasive testing, then you say it was much quicker and easier to use a debugger to disassemble modules and then you accept that it is possible that defendants did what [LzLabs’ SDM development lead] Mr. Jaeger described, then you say you cannot discount the possibility that other methods were used…,” Vanhegan said. 

“You have done all you can to discount what was said by Mr. Jaeger and come up with any other alternative that reflects badly on Winsopia and LzLabs.”

At the front and centre of IBM’s case is that LzLabs’s Winsopia testing subsidiary broke contractual terms by using its IBM mainframe to help LzLabs develop SDM. However, Vanhegan said that the SDM was functionally completed in 2013: that is, before the creation of Winsopia. 

Swanson agreed to this but in his report he wrote of “invasive and extensive testing” by the “defendants” during this time. Vanhegan put it to Swanson that at this time in 2013, Winsopia did not exist so Winsopia could not be considered in breach and nor could there be any “scrubbing failures” or failures to remove IBM code.

As well as LazLabs GMBH, IBM UK’s case is against LzLabs Limited, UK Winsopia and executives Mark Cresswell, Thilo Rockmann and serial entrepreneur John Moores, who bankrolled the company.

Vanhegan said: “So you didn’t mean to suggest Winsopia was involved?”

Swanson: “In initial implementation, no.” 

Vanhegan: “Nor LzLabs UK.”

Swanson: “I believe that’s true.”

Vanhegan: “Nor Mr. Moores.”

Swanson: “No.” 

Vanhegan: “Mr. Rockman.”

Swanson: “No.” 

Vanhegan: “Nor Cresswell.”

Swanson: “Correct.”

Vanhegan: “So LzLabs only…” 

Swanson: “Yes.” 

The end of the examination saw Vanhegan questioning Swanson’s credibility and independence, bringing to the boil what had long been simmering. 

LzLabs’ counsel (speaking rapidly in staccato almost throughout) suggested Swanson (providing long responses and frequently asking his interlocutor to slow down or repeat questions) could not be unbiased. 

This was variously, he alleged, because Swanson had worked almost exclusively in his career for IBM, most of the expert team that supported him in writing his report worked for IBM, he had not made clear aspects out of his area of expertise, he had no personal experience in writing for certain languages or programs, and because he focused largely on the claimant IBM’s evidence.

Vanhegan added that Swanson personally spent over 100 hours on the report but his expert team spent a great deal longer. Finally, he suggested that Swanson did not disclose the instructions he received, a requirement under law for expert reports

Swanson denied all of the above, maintaining he was fair and impartial and did not have any bias in favour of IBM. He relied on his expert team but maintained that he retained final say. Also, despite being retired since 2007, the alleged technical breaches are not novel, so his experience was still relevant. 

Repeating a mantra of his evidence, he also added: “I accept there was not adequate time to do analysis of all material.” 

The case continues on Monday when IBM’s second expert witness, Prof. Jon Weissman, takes the stand.

June 5, 2024: Mainframe expert goes deep into dev rights and wrongs as IBM v LzLabs case deepens

Defence counters IBM claims that its Swiss mainframe rival behaved negligently 

Described by the defence as an “IBM lifer”, mainframe expert witness for the claimant Michael Swanson continued his marathon testimony on day 21 of the IBM UK v LzLabs trial in London with details of the latter’s code development at the heart of matters.

In an examination that went deep into the guts of development techniques, defence lawyer Mark Vanhegan continued to press Swanson on details of how IBM’s Swiss rival came up with its Software Defined Mainframe (SDM) product that acts as an off-ramp from Big Blue for IBM mainframe workloads.

This involved detailed discussion of code fragments that would usually only be the talk of specialists and centred on the broader question of whether LzLabs and its mainframe software testing subsidiary Winsopia broke usage terms of an agreement with IBM.

LzLabs’ defence is that it made strenuous efforts to avoid copying IBM code and its Big Iron crown jewels, and that it had a legal right to examine code and interfaces based on the EU Software Directive. Further, any accidental infringements were quickly corrected.

Vanhegan argued that “any [code pointer] CSECT which IBM could have any propriety right to was removed from the customer application and not migrated to SDM.”

Swanson replied: “I accept that was the stated objective.”

Attention next turned to differences between IBM object code and SDM source code. Swanson only looked at the similarities between IBM code and SDM code and did not look at the many differences, Vanhegan suggested. Winsopia also never had a copy of IBM source code when developing SDM code but instead had object code. He argued that LzLabs designed an equivalent structure to IBM source code and deployed the results in the SDM. These were “semantically equivalent but not copies of each other”. 

Swanson agreed with this and agreed that the SDM developers had no choice but to do this. He added, however, that “there were aspects of IBM modules and runtimes that were learned and used in the SDM”, specifically referencing CICS, IBM’s communications software.

Vanhegan next turned his attention to a late change in Swanson’s report, only days before he took the stand, whereby a reference to similarity between IBM and LzLabs code was downgraded from being “strikingly similarity” to merely “similar”. Swanson also had changed his position that LzLabs had “highly likely made use of disassembled code by Winsopia”, he said.

Turning to the “scrubbing” process by which LzLabs attempted to cleanse its product of IBM copyright code, Swanson described scrubbing as a “laborious manual process”. 

Vanhegan suggested that while that may be true for customer applications, this was not the case for simple test programs where the process was very easy and highly accurate. Swanson agreed. 

Swanson also agreed he had not identified a failure in manual scrubbing of these test programs.

Swanson said LzLabs should have looked for eye-catching copyright and object code. Instead, Winsopia had used a reactive rather than a proactive process.

Vanhegan next asked Swanson to look into 18 instances of alleged scrubbing issues, suggesting these were specified because they are the “best examples IBM can come up with”. These, he added, must be compared in context of a vast number of clean examples. 

When it was put to Swanson that there could have been more context applied by his report, Swanson reverted to a familiar line: that “given time available and complexity, it is what it is”.

Vanhegan said: “In 99.999% of cases, scrubbing worked effectively.”

Swanson: “I can’t say that; I haven’t done the analysis.” 

Vanhegan dug in, suggesting that “as an IBM man, you didn’t want to [examine more deeply] because it goes against IBM’s case.”

Swanson replied: “I reject you calling me an IBM man. I didn’t think it was relevant to the case.” 

The case continues.

June 4, 2024: Scrubbing ‘failures’ and witness credibility: How Day 20 in the IBM v LzLabs case unfolded 

Expert witness quizzed over details of appropriate software development practices

On day 20 of this closely watched trial it was time to go deep as IBM expert witness Michael Swanson continued his long testimony, answering questions surrounding mainframe software development arcana and defending his own actions. 

This was a day of long probing into the scrubbing process by which the defendant says its Winsopia testing subsidiary cleansed code, using a tool it developed called CPX. Questioning turned to whether some transferring of IBM code was necessary. In his submitted report, Swanson referred to an example of this as a scrubbing failure but, for the defendant, LzLabs, counsel Mark Vanhegan said this was inappropriate and untrue.

Vanhegan next turned to what Swanson considered IBM proprietary material. “You suggested yesterday that … code generated by the compiler was IBM material…,” he said. 

Swanson replied: “There are IBM copybooks and macros that are clearly marked as proprietary. IBM also provides samples that may not be marked, and those I would assert are not proprietary and [there would be] no need to scrub those. Generated code, code that is put in by the compiler as part of the installation and language statement, evokes certain functions. That code, I had the impression, was IBM proprietary information.”

The vexed question of what could be legally shared is a leitmotif of this trial and Vanhegan put it to Swanson that the former IBM employee considered “every single transfer by Winsopia to have involved a distribution of some IBM proprietary materials, whether scrubbed or not”.

LZLabs counsel Vanhegan changed tack, noting that some 43 ‘historic’ CSECTs (or software pointers) were “irrelevant” to scrubbing failures because they could not have come from Winsopia’s mainframe. Why, he asked, did Swanson not make this clear in his report. Swanson admits he should have done, but had been limited by time constraints. 

IBM witness Swanson agreed, on occasion, with Vanhegan that he could have been clearer in his report, especially in instances where he agreed with other experts in joint statements. He also agreed that there are some aspects he should have added to his report that he did not – again, he said, because of time constraints.

It was put to Swanson several times that he did not verify points made by the defendant’s experts. Vanhegan suggested that Swanson was not a fair, impartial or independent expert based on his omission of material information in his reports and instances where he was not as clear as he should have been. 

Vanhegan asked repeatedly whether Swanson, who was assisted by a team of experts, had actually written the report. Swanson maintained consistently that he wrote it and stood by it. 

Vanhegan also said: “Every time that a defendant expert has done an analysis that undermines your suspicion that there has been a breach, you respond you didn’t have time to verify – did you not think it would have been productive to verify?” 

Vanhegan pressed on, suggesting that Swanson had been reluctant to agree with statements that “would be contrary to IBM’s case and your role”. 

Swanson responded: “I totally disagree with that … I have worked to provide a balanced, unbiased view and removed material from initial documents and considered documents provided by the defendants. Did I get to all of them? No.”

The case continues.

June 3, 2024: Expert witnesses’ take the stand

IBM expert and “Renaissance man” witness grilled on standard customer practices and modern host knowledge

On day 19 in the IBM v LzLabs trial, it was the turn of expert witnesses to step forward; the first 18 days were taken up largely by witnesses involved in events as they unfolded. If the first day of expert views is representative, much of the claim and counterclaim and disputatious jousting that has characterised the case so far will be replaced in this period by deep analysis.

The case is important in many ways but primarily, from a CIO perspective, in that LzLabs developed the Software Defined Mainframe (SDM) product for running IBM-compatible mainframe workloads without much of the hardware cost, staffing and maintenance challenges or software licensing ties to IBM itself. 

In that sense, it unpicks what is often seen as one of the great examples of customer “lock in” by which customers are tethered to a computing platform because of the overwhelming challenges migration would involve. In this case, IBM alleges that, in creating the SDM, LzLabs’ UK testing subsidiary and IBM mainframe customer Winsopia broke contractual terms.

The schedule calls for four expert experts to be called, two each for the claimant and defendant. The first witness called was at the behest of IBM: Michael Swanson, an IBM veteran who retired in 2007, described by the defendant in its skeleton summary document as an “IBM lifer” who LzLabs' lawyers said “appears to have had no real-world mainframe experience since 1999”.

LzLabs defence lawyer Mark Vanhegan described Swanson as “bit of a Renaissance man”, referring to his degrees in Chemistry and Biology, respectively after his IBM career had commenced and after retirement.

Swanson said that he had only become familiar with LzLabs and the SDM in 2023 when IBM commissioned him to write a report for this case, alongside a team of 11 experts. Swanson also testified that he had no working experience of modern mainframes specifically based on the z/OS operating system.

Vanhegan suggested Swanson had leant heavily on the other experts but Swanson denied this: “I wrote some of the initial draft perceptions and some were written by expert support team or legal team but that is all after information had been gathered and research done.”

Where material was sometimes removed or added “it was my call”, he claimed.

Although a draft of the report predated his involvement, Swanson had “substantially changed it to conform to his opinions and belief”.

He added: “I don’t know if there is a single instance where information in the report as it stands was unaltered from its initial draft”.

However, Swanson concurred that he had relied on research into the state of the IBM mainframe circa 2013 by David Stephens, expert for the defence, because he believed it was not necessary to verify findings and because of time constraints.

Throughout the case, LzLabs has attempted to normalise Winsopia’s activities and this line was present again in the following exchange. Vanhegan put it to Swanson that but was common practice for customers to disassemble their code to find bugs. Swanson disagreed saying this was “not a normal path”. 

Vanhegan: “I suggest, on the basis given by Mr. Stephens there was a common industry practice in 2013 to distribute compiled and load modules.” 

Swanson: “I have no reason to doubt that and I have not personally verified it.” 

Vanhegan: “One of the experts said this is the practice and trend at August 2013 and you don’t have an experience at that stage and you haven’t had the chance to verify so you have no reason to doubt Mr. Stephens.” 

Swanson: “That’s fair.”

Vanhegan: “IBM participated in this themselves and IBM also never expressed any objection to load modules or link-edited load modules.”

Swanson: “Correct.”

As a side note, David Stephens is the author of What On Earth is a Mainframe?

This case may have prompted a sale or two of that work.

The case continues.

May 30, 2024: Day 18: Mainframe experts weigh in as IBM-LzLabs trial meets a fork in the road

A quartet of Big Iron specialists gave testimony as factual witness stage draws to a close

Day 18 marked the commencement of the second half of this scheduled trial and the end of testimony given by factual witnesses, as a quartet of mainframe specialists gave their views and the case turned highly technical.

First, Alan Playford, a former Winsopia employee and consultant to the company, was cross-examined by counsel for the claimant IBM UK. Playford stated that he had worried that IBM would use any means to shut LzLabs down, including withdrawing its UK subsidiary Winsopia’s licence to use an IBM mainframe.

For IBM, counsel posited: “If IBM learned of the fact that Winsopia were using the mainframe to assist in the development of the SDM, IBM could well take steps to terminate the licence”. 

“I think that’s fair,” Playford replied, adding that LzLabs would take steps “to avoid advertising the connection”.

He added: “I had a fear knowing that we were helping to develop a competitor … that IBM would make it difficult.” 

But Playford also “had a suspicion that IBM may have known [of the LzLabs-Winsopia link] but it was not something that we knew for sure”.

Lawyers for the claimant IBM said the purpose was always to hide the connection and Playford reiterated that they were merely “avoiding advertising”.

“All of us knew that at some point the connection would be found; we couldn’t prevent that but we didn’t want to advertise that fact.”

Next under interrogation was Winsopia mainframe systems engineer Thomas Grieve, who was questioned about use of IBM code. 

“This is the way that software development works,” he said. “You write something, produce something and check how it works.”

He added: “It is inherent in the process that you are de-compiling third party material.”

The third witness of the day was another Winsopia mainframe system engineer, Kevin Lynch as IBM counsel asked the why he didn’t halt when he knew he was straying into IBM code.

“You were fully aware when analysing that you were out of the customer program and into the IBM code … There is no evidence that you stopped [on instruction from higher authority either]”.

“I just went where the analysis took me … It’s a standard technique,” Lynch said.

Finally, yet another Winsopia mainframe system engineer John Bray took the stand and was presented with messages that suggested the software scrubbing process wasn’t working fully effectively.

“I think the process is not preventing IBM material getting through [sad face emoji],” he had written.

The case continues. 

May 29, 2024: Day 17: LzLabs holds to separation lines under IBM questioning

IBM persists in allegations that mainframe rival must have shared Big Iron know-how

As the trial between IBM UK and LzLabs reached its midway point, LzLabs senior executive Mark Cresswell, just as witnesses before him, maintained the line that it was never important that IBM discover the connection between UK subsidiary Winsopia and its Swiss host. The choice to market everything through LzLabs was, according to Cresswell, “to be as clear as possible to whoever was listening because it was a marketing message – one has to be crisp – and the SDM [IBM mainframe alternative] was an LzLabs product.” 

That statement pretty well summed up one of the key aspects of the case, which has often reverted to the theme of whether IBM knew at an early stage that LzLabs and its Winsopia unit were connected.

Cresswell explained that, in 2020, LzLabs had reached such a level of development with the SDM that it was considering getting rid of the Winsopia mainframe entirely. One option was to fold Winsopia as a business unit into LzLabs and IBM’s ICA contractual agreement and the EU Software Directive had given “historical protection in our approach for development”.

In May 2020, LzLabs UK was folded into Winsopia.

In a demonstration video filmed in 2016 at Winsopia, Cresswell took care to ensure that no Winsopia name or logo was in the video, it was suggested. Claimants said this was because LzLabs wanted to conceal its link from Winsopia but Cresswell said repeatedly that this was marketing rationale in pursuit of clarity for customers. 

Cresswell continuously downplayed his role and said he didn’t approve changes to the Code of Conduct that governed separation of duties between LzLabs that ensured laws were not broken.

Cresswell said: “In the same way, perhaps, that I have the ultimate say in how financial filings …  you wouldn’t expect me to look over the shoulders of our accounts. I’m not skilled in that area … It would have been careless of me to interject myself in this process, particularly when we have members of staff who have more experience.”

Cresswell said he was not aware of any LzLabs employees having access to the Winsopia-licensed mainframe but the claimant pointed out that 14 non Winsopia people were inadvertently given access to the mainframe room. Cresswell responded that access to the room was not the same as access to the mainframe.

Cresswell said Winsopia, in his experience, ran “a tight ship”. He admitted that he was “ecstatic” about the lifting of restrictions in the Code of Conduct, which would speed up development of the SDM. However, he said, “I remember, during my first few months, being taken aback by the amount of attention given to this [governance] process.”

Cresswell said that he wanted LzLabs to succeed but “not at all cost” and that he “was acting in good faith at all times”.

May 28, 2024: Day #16: La-La Land, the clean-room Guinness Book of Records and odious regimens: IBM v LzLabs gets more feisty

Colourful exchanges between IBM lawyers and veteran entrepreneur John Moores mark day 16 of the far-reaching mainframe case

The trial involving IBM UK and its far smaller mainframe rival LzLabs is well into its stride now, and, while the toasty environment has been a constant that has defied various attempts to cool down the London courtroom, there is no doubt that the trial is warming up to a possible peak. Presiding judge Mrs Justice O’Farrell will soon need to make decisions that impact, in some significant way, the future of enterprise computing. 

Regular readers will know that this case has focused on two areas.

First, pragmatically and specifically, on whether LzLabs’ UK subsidiary Winsopia broke terms of being a mainframe user under IBM’s ICA contractual terms. And, second, more broadly, on whether LzLabs attempted to hitch a ride on the many-decades success of IBM’s mainframe technology.

IBM has attempted to show that LzLabs and Winsopia worked in cahoots, sharing technical knowledge of the workings of IBM’s mainframe beyond what is allowable. And, as witnesses called have become progressively more important, the case has taken on elements of a courtroom drama with the latest exchanges particularly feisty and colourful. 

On day 16, Nicholas Saunders KC for the claimant IBM and US entrepreneur John Moores jousted in more lively exchanges. Many of these concerned how much Moores knew about the development of the LzLabs’ Software Defined Mainframe (SDM) product he funded. That product threatens IBM’s mainframe revenue stream by being able to handle its mainframe workloads on commodity servers or in the cloud. 

Moores strongly asserted that he was not that involved in the development process and disputed that his involvement in a previous case, via another attempted mainframe contender called Neon Enterprise Software, should have made it clear that Lz’s tactics were illegal.

Moores was questioned about his role in hiring a number of people but said he “did as much as I could, [but I] don’t recall if I was the first to recruit [them]”. He had signed off hires as a “recruiting officer” on one of his emails but said this was merely a light-hearted comment.

Talk turned to the Code of Conduct [‘Code’ from here], put in place to create a clear separation of duties between Winsopia and LzLabs. Moores stated that he had never read any version of the Code and was pushed by Saunders on this:

“You sent through emails about the strictness of the Code … You must have had an awareness of what it applied/implied; it was important to you that you had fair rules around who could communicate with who.”

Saunders asserted that Moores had communications with both Lz and Winsopia. Moores replied: “I had periodically some communications, but it would be a gross exaggeration to suggest that there was a lot … [That’s] misleading.”

Moores asserted that he had a general understanding of the Code and that he adhered to it. Saunders persisted: “That understanding was not such that you didn’t inadvertently breach the rules from time to time.” 

Discussion next turned to the development process where Moores stated he was unhappy about the pace of communicating with Winsopia and was frustrated about the so-called ‘Chinese Wall’ by which developers create a gulf between areas of specialist knowledge.

Developers at Lz had to make a request through a lawyer in order to interact with Winsopia for trivial matters, he contended but he  “wanted the Code to be more efficient.” And to the notion that he applied pressure on employees such as LzLabs CEO Thilo Rockmann, he said that “pressure is too strong a term”.

Moores also stated he never read IBM’s ICA contract on computer usage at LzLabs, nor at Neon. “Few people that work with IBM computers have actually read the ICA … [that’s] for lawyers,” he said. Instead, “I made sure that competent lawyers read all documents, including the ICA.”

Saunders harped to a favoured claimant theme, suggesting that Moores must have been aware of the ICA contents to a degree, given the Neon litigation.

“I’m not a lawyer but I’m confident that the professionals understood the Code that had been written and that they would scrupulously adhere to it,” Moores said.

However, he added that he had become convinced that maintaining the LzLabs/Winsopia Chinese Wall had become “an absurd waste of time and money, waste of programmer resources in maintaining the clean room – undoubtedly in the Guinness Book of Records for longest running clean room”. (In LzLabs submissions, the company claims to have spent well over $300m on R&D for SDM.)

Moores agreed that he was “frequently frustrated with Lz” but denied that the effect of that frustration was an amended Code. The development process as related to the LzLabs-Winsopia relationship “was inherently wildly inefficient and that was just the price to pay for a clean room”, he added. 

But, he claimed, Lz had continued to maintain strict processes. “We kept records to show that the SDM had been properly developed in the most stringent environment … Don’t imagine a more odious regimen to develop software in the history of mankind.”

Further, Moores said: “As you are aware, IBM has unlimited financial resources and a long history of being very aggressive. Many companies have fallen at the wayside at the efforts of IBM … The entire purpose … [was] to prevent a bogus claim by IBM about copyright infringement.”

Saunders said that, as the ultimate owner of LzLabs, people would naturally demur to his wishes. And, as a result of changes to the Code, “operational separation was undermined between Winsopia and LzLabs … By making those changes there was a substantial increase in the threat there might be breaches of the ICA … that was risk you were preparing to run because your main concern was getting development of the SDM completed and you weren’t concerned with that breach [of the ICA].” 

Moores said this was utterly untrue.

 Later, Moores suggested to Saunders that he was “grossly misinterpreting my role as an owner … I didn’t direct anything and it’s annoying as a witness when you repeatedly use terms that suggest things that were not the case, I didn’t direct anything … I’ll accept ‘encourage’.” 

While Saunders suggested that Moores was in “the weeds”, Moores compared himself to a pilot: “I’m flying at 50,000 feet”.

Moores said: “At all times, LzLabs was quite concerned about not using or transporting IBM material or other products”. 

Instead, the focus was on dealing with customer problems and all work that was done with scrubbing IBM code was about “find[ing] problems in IBM software [and] stopping slow performance for the customer”.

Moores was told by the judge to answer the questions only rather than making general statements.

“I feel like I’m in La-La Land,” Moores said about the discussions about the ICA and the goal of Winsopia. “You’re drawing up a conspiracy painting me as a tech genius who knew everything … That’s not the case.”

The case continues.

May 23, 2024: Day #15: The Secret Formula: Tech entrepreneur John Moores testifies on the quest for LzLabs’ interoperability with IBM mainframes

In the latest gripping instalment of the IBM v LzLabs case, the man who bankrolled the project takes the stand

Legal complexities surrounding mainframe computing may not electrify all of the world but for those that understand the massive stakes here, this was another fascinating day in the trial of IBM UK versus its Swiss rival LzLabs as serial technology entrepreneur John Moores took the stand.

Moores was a co-founder of enterprise software giant BMC (in which his name stood for the ‘M’) and ServiceNow. He has been better known in other circles as the one-time owner of Major League Baseball’s San Diego Padres and as a potential bidder for English Premier League football involvement. Closer to home in terms of this case, he also funded LzLabs with the aim of providing an off-ramp for IBM mainframe users to run workloads on commodity servers and in cloud datacentres. That is no small investment and LzLabs’ skeleton statement suggests over 300 million Swiss Francs ($327m) went into R&D for the project.

This was not Moores’ first bid to upset the IBM gravy train as a previous effort, Neon Enterprise Software, had hit the buffers following another legal case in 2011. On day 15 of this latest case, heard in London’s Technology and Construction Court, Moores was asked to recall his involvement in LzLabs.

IBM’s lawyers began by asking about Swisscom, the lighthouse win for LzLabs as the first company to replace dependence on IBM Big Iron for Lz’s Software Defined Mainframe (SDM) alternative. Moores agreed that the win was important but said he had had no direct involvement. IBM had “a perfect monopoly on mainframe computers and software”, he added.

Moores said that he felt emboldened in the LzLabs project by the EU Software Directive where the “clear purpose … was to tell IBM they couldn’t interfere with companies making their programs interoperable”.

IBM lawyer Nicholas Saunders clashed with Moores when the former suggested that the latter had instructed developers to “crack” IBM’s mainframe code.

“It’s incorrect that I would have told him to crack it,” he said. “I never told anyone to crack anything.”

There was more testiness next when Moores was asked why he had been personally named in the Neon trial. “Ask your client,” Moores replied. When Saunders persisted, Moores insisted that this had been “a crazy overreach”.

Moores went on to compare accessing the interfaces to IBM mainframe code to understanding the formula underlying Coca-Cola.

Attention next turned to Moores’ emails and the loss of Gmail messages IBM lawyers had requested to see. Moores said it was “not my customary practice to send emails to anyone at LzLabs not through my LzLabs email”. 

When Saunders suggested an example to the contrary, Moores said emails did not contain references to LzLabs’ SDM product.

And when Saunders suggested that the disappearance of emails was deliberate, Moores asserted that this was “offensive” and “utterly incorrect”.

Next, Saunders attempted to position Moores as a “prime mover of the SDM” but Moores demurred, saying he was merely an investor with an outdated knowledge of the tech sphere.

It was suggested that Moores was unhappy at slow progress with the project and there was laughter when asked how this manifested itself.

“Generally, I would say ‘I’m not happy’,” Moores said.

Persisting gamely, Saunders suggested staff would inevitably “jump when you say ‘jump’”.

Moores again disagreed, suggesting that owners who are less technical than company leaders have a limited impact on projects. Well over half of SDM development was outside his realm of knowledge, he added.

Neither did IBM’s legal team progress far when they returned to the familiar refrain that LzLabs had attempted to conceal its relationship with UK subsidiary Winsopia. “I think that’s silly,” Moores said.

Moores also earned laughter when asked about the emails he had seen in a previous career as an IBM salesman.

“There were no emails when I was a salesman,” he said.

In a stuffy atmosphere where thermal controls have been a constant environmental backdrop, Moores earned yet more guffaws when told he could remove his jacket while testifying. 

A regular visitor to the trial this far, “It’s considerably cooler up here than in the cheap seats,” he deadpanned.

The case continues.

May 22, 2024: Day 14: Mainframe case warms up as tempers fray

Heated exchanges flare up as IBM casts doubt on rival’s development practices and adherence to code of conduct

The London court case between IBM UK and its Swiss mainframe workloads challenger LzLabs warmed up today as Keith Rastall, founder and later general manager of LzLabs’ UK subsidiary Winsopia, took to the witness box. IBM alleges that its rival broke terms of the ICA contractual agreement and exchanges became somewhat heated as LzLabs’ lawyers objected to some lines of questions.

Rastall recalled that in February 2013 he was invited to Kentucky to meet technology entrepreneur John Moores and Thilo Rockman, who would go on to become LzLabs CEO. At the instigation of Moores, the trio discussed the creation of a new company that would acquire a mainframe and provide services to LzLabs. That led to the creation of Winsopia, which would later be sold to its sole client, LzLabs.

Always quick to suggest uncomfortably tight links between the work of LzLabs and Winsopia, IBM’s legal team wasted no time in going on the offensive.

“Based on documents we have seen, I suggest that, first, LzLabs needed access to a mainframe but wanted to procure it from an independent company to avoid IBM learning that mainframe would be used for LzLabs purposes,” stated Fred Hobson KC for the claimant. 

“Second, you were approached by Mr. Moores and his team and asked to set up a company for that purpose, and, third, it was never intended that Winsopia would ever be used for anything other than a service company for LzLabs. And finally, when Winsopia was incorporated, it was acting at the bequest of LzLabs and John Moores.”

“I disagree,” Rastall replied.

Hobson suggested that Rastall was concerned about “IBM learning that Winsopia, its licensee, was using the mainframe to assist in the development of a competitive product, and IBM could then take steps to prevent the development of [LzLabs’ competitive product] SDM”. 

Rastall agreed with this statement and also with Hobson’s assertion that “the fact that Winsopia was assisting in development of SDM is not something you wanted IBM to know”. 

However, when Hobson added that the “fact that Winsopia was owned by LzLabs was not a fact you wanted IBM to know”, Rastall said it was inevitable that IBM would realise the link. 

Throughout the trial, LzLabs has said consistently that it did not attempt to hide the connection between the organisations.

Hobson said, “This was a deliberate policy put in place to conceal the link between Winsopia and LzLabs to hide the breaches of [IBM licensing agreement] the ICA”. Rastall denied this and added that when discussions took place over how Winsopia staff could describe themselves in 2016, he was “absolutely certain” that IBM knew of the link and there was no attempted subterfuge.

Rastall himself had signed the ICA and said he had a “reasonable” comprehension of it. It would be unlikely that anyone but a lawyer could claim to have a comprehensive understanding of it, he added, but he had reviewed the document with care and had taken legal advice from Clifford Chance.

Rastall said he had believed until very recently that people with visitor badges at Winsopia were not granted access to the mainframe room but now recognised that this was not true. 

“So, the entire time you were supervising at Winsopia you assumed that access to the mainframe was not possible with a visitor pass and now you know you were proceeding with a false assumption?” Hobson asked.

Rastall conceded this was the case but added the caveat that the impact was “minimal”.

LzLabs developers were required under the firm’s code of conduct to take notes of any face-to-face discussions with Winsopia employees when they were at the latter’s offices. No notes have been disclosed in this case and Hobson contended that this is because this rule was never actually followed. He identified a series of LzLabs developers who went to Winsopia and asked for each if Rastall saw notes taken by these developers. To this, his respective responses indicated he didn’t recall or doubted that notes had been taken. 

In his witness statement, Rastall said he was sure that Paul Knight, IBM account manager and the customer account manager for Winsopia between 2013 and 2108, would have known about the LzLabs and Winsopia connection, as well as exactly the type of work Winsopia was doing, because he visisted Winsopia and would have seen employees wearing LzLabs merchandise. Hobson suggested that it was “entirely implausible” that he could remember anyone wearing LzLabs clothing in that period. Rastall said that he clearly recalled Knight and he drinking coffee when Rastall would have sipped from an LzLabs mug. He also said that had specific conversations with Knight about what Winsopia was doing.

Turning to another recurring theme of the trial, scrubbing, Rastall disagreed with the suggestion that LzLabs’ scrubbing tool was “inherently unreliable”. Rastall agreed however that Winsopia did pass load modules to LzLabs where there was a risk of them containing IBM code.

Rastall also sent emails to Moores with non-scrubbed or unredacted IBM load modules which essentially circumvented the Discovery Request (DR) system it had in place. This happened on three occasions shown to the court. Hobson asked if Rastall considered whether the DR system applied to Moores. He replied that he didn’t. Hobson suggested this was a “significant oversight” on Rastall’s part. Rastall accepted this but said these were rare occurrences.

The case continues.

May 21, 2024: Day #13: IBM queries separation of duties at LzLabs and its UK subsidiary

Big Blue lawyers stick to familiar refrain in mainframe case

Throughout the high-stakes trial, the legal team for claimant IBM has attempted to punch holes in LzLabs’ argument that it maintained strict separation of duties between itself and Winsopia, the Swiss firm’s UK subsidiary. Day 13 was no exception as the claimant continued to chip away at witnesses in a bid to prove that, in developing its mainframe software, LzLabs broke IBM’s ICA licensing agreement.

Today was the turn of Christopher Palmer to receive this persistent treatment. Palmer, a software engineer at Winsopia from 2013 through 2019, rebuffed suggestions that the relationship between LzLabs and Winsopia caused leaks and exchanges of information.

Responding so quietly that there were regular requests for him to speak up, Palmer said that LzLabs’ code of conduct stated that employees with access to the Winsopia mainframe could not work on LzLabs’ Software Defined Mainframe (SDM). 

The development of SDM was a breakthrough because it provides IBM customers with the ability to run mission-critical workloads without the costs of deploying and operating mainframe hardware. 

For IBM, Matthew Lavy KC put it to Palmer that he had a close relationship with John Moores, the tech entrepreneur who funded LzLabs, to the extent that Palmer considered himself an employee of Moores rather than the company. Palmer confirmed he had sent Moores emails without cc’ing his line manager but said that, although Moores was friendly, there was no possibility of this.

Another line that IBM has repeatedly pushed is the notion that its rival’s software scrubbing process was weak.  Lavy put it to Palmer that the improvements to his scrubbing process, such as the use of wildcards, were suggested to him. The wildcard enhancement was something that a lot of witnesses said would improve the effectiveness of the scrubbing technique to ensure no IBM code would cross over. Palmer said today that despite being made aware of this he never made any changes in relation to wildcards and while in hindsight it might have been a good thing to use them, it wasn’t done and was never a priority. 

Palmer left Winsopia to work at Texas Wormhole, a company that provided development services to LzLabs. Lavy suggested that, while Palmer didn’t work on SDM development, there was a possibility that he hadn’t been able to “flush it all out of his mind”.  

A thermometer displayed 23 degrees Celsius on another warm day in the London courtroom that, in terms of evidence given, also turned out to be a rather dry one. The case continues although some witnesses listed will not now be required to give evidence in order for the trial to fit the scheduled nine-week time window.

The case continues.

May 20, 2024: Day #12: IBM and LzLabs continue to tussle over heart of mainframe future

London case may dictate the next-generation data centre 

IBM and its far smaller rival LzLabs continued to dispute on day 12 of their trial that may control much of the future of enterprise data centre computing.

The first key witness was David Bond, a senior software developer at Switzerland-headquartered LzLabs, who was initially employed at another company, Texas Wormhole, from 2012 to 2013. Bond moved to LzLabs and worked on the LzLabs’ SDM software alternative to IBM mainframe computing from 2013. 

The claimant, IBM, continued to question the effectiveness of LzLabs’ software scrubbing and clean-room processes, suggesting there was no true separation between LzLabs and its subsidiary Winsopia in the developing of the SDM.

The next witness, mainframe veteran Brad Taylor, described how he had been given the special mainframe permissions, aka ‘Brad’s Nook’. The computer was physically locked in Switzerland and before his first secondment in 2014, Taylor said he probably accessed the ‘Nook’ over 500 times with ‘airlock’ privileges. 

IBM counsel suggested Taylor was “frustrated” by the Discovery Request system with LzLabs subsidiary, Winsopia. Taylor said. “That’s putting words in my mouth … It was painful but necessary – it was slower than if I did it myself, but it was necessary.”

Taylor agreed that the system was slow and the Winsopia employees were not skilled enough. This slowed down his work and made it impossible to reach a certain milestone set up by LzLabs’ funding source John Moores. Therefore, it was decided that Taylor should go on secondment to Winsopia. But there were strict controls on usage, Taylor, said.

The case continues.

May 16, 2024: Day #11: IBM probes test of ethics walls used to build LzLabs’ mainframe software

Claimant queries resilience of rival’s separation of duties in developing its alternative technology to IBM Big Iron 

In the latest twist to the trial concerning legal development of alternatives to IBM’s mainframe workload business, the IT giant cast doubt on its smaller rival LzLabs’ processes for keeping developers with conflicting duties apart. 

The separation process is often known as “Chinese wall/s” although many prefer the term “ethics wall/s” today.* The process has long been a staple topic of debate in software development, for example, in the 1990s when it was widely pondered whether Microsoft application devs had a leg up through knowledge of the same company’s Windows operating system.

On day 11 of the trial, that notion was aired again when Christian Wehrli took the stand. Wehrli is a former LzLabs and, indeed, IBM employee but today works for neither concern. He gave evidence about the clean room process in which LzLabs says it developed its Software Defined Mainframe (SDM). SDM is a product that lets customers run mainframe workloads without access to an IBM mainframe, thereby threatening IBM’s income from its long-running but still hugely important business area.

For IBM, Nicholas Saunders KC once again began by taking aim at a defence witness’s credibility, saying Wehrli wasn’t an appropriate person to comment on processes at LzLabs’ Winsopia UK subsidiary as he had not been employed by Winsopia and hadn’t overseen its clean room operations on a regular basis.

Saunders tried to break down the claim by LzLabs that Winsopia and LzLabs maintained separate physical locations. Wehrli said that LzLabs could use a VPN, which allowed its employees to access the server from anywhere and confirmed that no Winsopia employee had access to this. LzLabs AND Winsopia also operated different discovery systems, he added.

At one juncture in 2015, LzLabs UK and Winsopia shared an office but were separated by a (physical) wall, Werhli said under further questioning. 

This led Saunders to suggest:

“This idea that the offices operated entirely separately was not reality … [Staff] were mixing quite fluently … getting coffee and chatting away to each other.”

This was rebuffed by Wehrli:

“As far as I’m aware, everyone was taking this very seriously and we knew there was a certain risk involved with proximity between the parties, which was discussed … with legal counsel but we decided it was a risk that could be [managed] and everyone was extensively briefed as to what they could do and what they couldn’t. I am not aware of anyone who did not take this seriously and every time I visited the premises, they all knew what they could and couldn’t do.”

It was put to Wehrli that he had written an email saying, “We don’t want to reveal the Winsopia name and terminate the licence for [the mainframe].”. Saunders said that Wehrli did not want to reveal, and in fact purposely concealed, the Winsopia and LzLabs connection because he knew that Winsopia had breached the terms of its mainframe licensing agreement. The idea that LzLabs concealed links between it and its subsidiary has been a recurring theme of the trial so far.

Wehrli said he did not want to “advertise” the connection between the two companies, even though it was public knowledge and he thought that IBM might have already known about it. However, he was concerned about IBM becoming aggressive and possibly terminating the license with Winsopia if it felt threatened by the SDM product. Wehrli added that he didn’t want to confuse customers.

“My fear was generally that they would strike a competitive product, taking the licence away,” Wehrli said. “I’m not a lawyer. I don’t know all the possibilities, but I was genuinely concerned.” 

Saunders suggested that Wehrli was aware of a potential breach of the ICA mainframe usage licence. Wehrli denied this.

This has been a trial studded with many tetchy exchanges and Saunders appeared to be in a particularly feisty mood, suggesting that Wehrli was providing “stock answers” that were irrelevant to many questions he was asked. The claimant’s lawyer asked numerous times if someone had told him to “give these speeches”. Wehrli was adamant that nobody had coached him and everything he was saying was his own words. 

The next witness, LzLabs senior software developer David Bond, was quizzed about what the claimant’s lawyers considered a gap between his witness statements saying the scrubbing process at Winsopia was highly effective and emails he had sent suggesting improvements. Bond said there were differences in quality of the scrubbing process depending on development language and that he had made suggestions for enhancing processes.

The case continues.

May 15, 2024: Day #10: LzLabs CTO defends development processes under IBM questioning

Mainframe veteran Jan Jaeger backs up coding processes for its competing SDM technology 

Jan Jaeger, LzLabs chief technology officer and chief software architect, continued to give evidence on day 10 of the trial in which IBM UK alleges that its mainframe rival LzLabs failed to comply with the former’s usage contract, the ICA, giving the Swiss company the chance to create an alternative to IBM’s long-running host server line. For the claimant, Nicholas Saunders KC attempted continually to punch holes in the credibility of the witness, suggesting his statements were misleading or plain incorrect.

Jaeger, a tall, softly spoken mainframe veteran, was not this time asked to raise his voice for the better understanding of the London courtroom as he had been the previous day. However, he continued to provide deep insights into LzLabs’ processes as he countered suggestions put forward by IBM counsel. 

Jaeger agreed that he was not intimate with the ICA per se but said he had always been confident that LzLabs was acting legally based on his understanding of the licensing terms and conditions of specific programs and a knowledge of the EU Software Directive. The Directive broadly supports interoperability between vendors. 

For IBM, Saunders also suggested that Jaeger had only partial control over the DR (Discovery Request) system that governed LzLabs’ internal collaborative attempts to understand mainframe interfaces. As on the previous day, Jaeger responded that it was impractical to look at every line of code, adding that he oversaw a team conducting a multi-step process. When Saunders posited that a junior member of staff might not feel confident enough to admit to an error, Jaeger gave a response straight out of the professional developer’s playbook:  

“People make errors; that’s not bad,” he said. “What is bad is not correcting the error. That’s how we operate at LzLabs.”

Probed on the quality of LzLabs’ scrubbing processes to remove and replace “stubs” (fragments of code that as pointers), Jaeger conceded that the process was not perfect but suggested that “over-scrubbing” also has issues. “There is no perfect solution,” he said. “We were doing the best we [could] and I think we did a good job.”

At the heart of the case is whether LzLabs and its UK subsidiary Winsopia broke terms of the ICA in their use of an IBM mainframe acquired by Winsopia. Saunders suggested that LzLabs could not have created its SDM product for running mainframe workloads on commodity servers without access to the IBM server hardware. Jaeger demurred but acknowledged that having a mainframe accelerated some areas. Not having to wait for load modules did “speed up the process”, for example, Jaeger said.

The case continues with further witnesses for the defence scheduled for the next days. 

May 14, 2024: Day #9: LzLabs CEO completes epic session amid mainframe contract breach claims 

Thilo Rockmann finally stood down after two-day testimony

As the IBM-LzLabs trial reached its projected quarter-way point, key witnesses for the defendant continued to bat away suggestions that it had breached IBM’s ICA customer agreement in developing its Software Defined Mainframe (SDM) product. SDM is designed to let mainframe users run workloads in modern environments without the costs or overheads of Big Blue and ‘Big Iron’ hardware. The morning of the trial notably saw the close of a marathon session within the trial as claimant lawyer Nicholas Saunders KC wrapped up questioning of LzLabs CEO Thilo Rockmann that had run for over two days. 

In a packed court of about 50 people, continued issues with heat management saw the London courtroom strewn with water bottles and hear the hum of cooling fans. There was heat too in some lines of claimant questioning.

Saunders repeatedly suggested that LzLabs management must have anticipated legal challenges from IBM.

“You didn’t seriously expect that IBM would be unaware of intellectual property infringements?” he asked Rockmann.

Rockmann returned to the theme of the defence which is that customer applications were not impacted by ICA and that LzLabs had only been interested in legally understanding critical interfaces on the mainframe. The comparison the defendant makes for its SDM product is with an international power adapter designed to work across devices and outlets.

Saunders turned to IBM’s ICA auditing provision, stating, “It must have been absolutely obvious you were going to get an audit.” Later he asked, “What did you think IBM was going to do? You just sat there passively wondering?”

Rockmann said that he had not expected a sudden escalation of threats and when he had received a letter from IBM suggesting there had been a “material breach” of the ICA he had been “shocked and angry”. He said: “I never, ever thought a platform vendor would claim rights to a customer application.”

Saunders alleged that Rockmann and LzLabs had tried “to be as obstructive as possible … You chose not to supply a single thing [requested by IBM]”. But Rockmann said there had been a “lack of engagement on material matters” and questions had arisen over confidentiality. LzLabs and its Winsopia subsidiary had always tried to comply with the audit process, he claimed.

Saunders concluded his cross-examination of Rockmann, saying, “All of you were happy to put in place measures that allowed breaches to happen.” 

But Rockmann denied this. “We wanted SDM to succeed [but] I wouldn’t say at all costs, at all,” he said.  Instead, he said, LzLabs had always acted fairly with the aim of helping customers to modernise their infrastructure. The relationship with Winsopia that IBM alleges was concealed “could have easily been found by anyone [as] public information on the internet”.

“That’s what competition is all about,” he said. “It’s normal [and] very attractive to customers … We were trying to do the right thing. We are human beings and mistakes happen [but there was] no breach or IP infringement. I did act in good faith … I don’t think there was anything misleading in any form.”

The next witness called was a mainframe veteran of over 40 years.  Jan Jaeger, CTO and chief software architect, LzLabs, was questioned over how he and other experts were recruited to LzLabs and on technical matters but his early testimony was leavened by a humorous exchange when it was suggested that he had submitted four statements. Jaeger appeared confused before saying, ‘I thought you had said false statements…’” “We’re not getting there yet,” Saunders responded, laughing.  

It was put to Jaeger that he must have expected legal action but he explained that although “people were speculating”, the example of another company suggested that sustainable mainframe-sector businesses could exist:

“Micro Focus has been doing it for 30 years and they’re still here,” he noted, referring to a UK-founded software veteran that is now part of OpenText and is best known for helping customers to modernise mainframe workloads.

Saunders and Jaeger jousted over definitions of reverse engineering and code interfaces and when Saunders suggested that compiler listings are for debugging, he was immediately rebuffed with Jaeger pointing out that many banks use them explicitly for audit purposes.

When Saunders turned his attention to compliance with the LzLabs Code of Conduct, Jaeger said that the managed process through which LzLabs communicated about mainframe interfaces with its Winsopia subsidiary was effectively monitored by Clifford Chance lawyer Daniel Hedley in the Discovery Request system.

The case continues with more defence witnesses working for LzLabs scheduled to be called next.

May 13, 2024: Day #8: LzLabs CEO rebuffs IBM suggestions of backdoor agreements and secret liaisons

Continuing his epic stint on the stand, LzLabs’ Thilo Rockmann defended the Swiss company’s mainframe R&D agenda

In the fall-out from another high-profile technology court case that took place in the early 1990s, Intel famously referred to microprocessor rival as “the Milli Vanilli of semiconductors”, a quipping reference to the pop duet who didn’t sing on a hit marketed under their name. IBM UK lawyers haven’t quite gone so far but they have left zero doubt that they believe mainframe computing workload rival LzLabs broke a key contract relating to mainframe usage to develop its rival product.

In another long day in court for Thilo Rockmann, CEO of the Swiss-based company, the court heard repeated allegations by the claimant’s Nicholas Saunders KC that LzLabs had hidden its relationship with Winsopia to swap information that led to the development of LzLab’s Software Defined Mainframe (SDM) product. The allegation is key because Winsopia, a UK-based subsidiary of LzLabs, is the organisation that acquired the mainframe in question and IBM says it broke the core ICA contract relating to usage terms and conditions.

Rockmann was at pains to suggest that a code of conduct was monitored and adhered to throughout SDM development, ensuring strict separation of Winsopia’s QA and LzLabs’s R&D areas of specialisation. 

However, Saunders suggested that lines between development, discovery and testing became blurred. Visits to Winsopia’s Farnborough, Hants. offices by LzLabs left the door open to LzLabs staff being aided in efforts to reverse engineer the IBM software environment, Saunders suggested. 

Arrangements meant it was possible for LzLabs staff to “slide over and have a cursory look over the shoulder” of Winsopia employees, he said. “There were significant departures from what was arranged…this was no longer a separation”, he added, and “valuable knowledge could be shared over the course of a short discussion”.

Rockmann said there had been no connivance and no collaboration that touched on development of the SDM or impacted the ICA.

While Rockmann denied that there had ever been a plan to obscure the relationship, Saunders pushed on with his well-worn line: 

“You were concerned about making sure that IBM were not tipped off about a connection between Winsopia and LzLabs,” he alleged.

Rockmann also defended LzLabs on the point that employees had email addresses for both LzLabs and Winsopia domains. These were to support separations of duties and to provide focus, he said. 

Earlier in the day, Saunders pursued another familiar refrain, suggesting that LzLabs “scrubbing” of software had been inefficient. When customers compiled progams, the IBM load module would automatically introduce IBM fragmentary code “stubs” into the application. IBM’s counsel suggested that LzLabs engineers attempted to replace these stubs with scrubbing in order to avoid transgressing the ICA. Rockmann said that it was customers’ own applications that contained these stubs and so outside LzLabs’ legal domain.

The case continues and indications are that Rockmann will be queried for about half an hour in the following morning session.

May 9, 2024: Day #7: IBM goes on attack, challenging LzLabs’ ‘Chinese walls’ in making its mainframe workload rival

Claimant’s lawyers double-down on Swiss rival’s development processes 

Day seven of the IBM UK claim against mainframe workload rival LzLabs saw IBM lawyer Nicholas Saunders KC take his turn to examine witnesses, beginning with an exhaustive questioning of the Swiss company’s CEO, Thilo Rockmann. The line of attack focused on whether LzLabs had acted inappropriately and hitched a “free ride” on IBM technology in developing its Software Defined Mainframe product that can run mainframe workloads on independent hardware and operating system.

In a session lasting all day, Rockmann was quizzed over his relationship with John Moores, who bankrolled the company; his experiences at a previous company, Neon Enterprise Software; and the extent to which the Swiss company adhered to its internal code of conduct. Notably, Saunders focused on whether ‘Chinese walls’ (that is, strict segregation of developer staff acting ethically) had been effectively deployed to maintain compliance with IBM’s ICA customer contract. 

The session began with Saunders going back to Rockmann’s time at Neon, a previous Moores-owned company that created zPrime, a way to offload IBM mainframe processes at lower cost. Neon collapsed in 2011 in the wake of another trial with IBM. Later, Moores asked Rockmann to get involved in creating  a new project that became LzLabs, appointing him as CEO and chairman.

Pressed on details of the Neon collapse and settlement, Rockmann said he was unaware of details. “I was fired,” he said. “Neon was history for me.”

When Rockmann was later in receipt of letters sent by IBM special projects executive Mark Anzani, Saunders suggested: “You knew very well what was going on. You’d been through all this before.” Rockmann responded: “You’re trying to combine things that are not related ... The contents [of the respective letters] were very different.”

Rockmann went on to say that there had been no widespread expectation of IBM litigation against LzLabs but he stressed there had been no relaxation of a code of conduct established to maintain order in clean-room development. Notably, there had been separation between software development and QA testing for optimal efficiency, he claimed. 

Rockmann said LzLabs had not prepared to brief customers on potential IBM allegations and actions but said it would explain product development processes when asked. Despite this, he conceded that some customers had “disappeared” after receiving communications from IBM.

Saunders also pressed Rockmann on its UK subsidiary Winsopia’s acquisition of a mainframe, suggesting that LzLabs had concealed its relationship with the Farnborough, Hants.-based company. He next returned to Rockmann’s relationship with Moores, a serial entrepreneur who co-founded systems management software giant BMC Software. (Moores later also found fame as the former owner of baseball team San Diego Padres, as well as via links with moves into English football.)

Rockmann conceded that Moores “could be demanding” and “frustrated”, manifesting this in emails and calls. “He was the ultimate investor and, understandably, he wanted to see a return on his investment,” Rockmann said. Asked about Moores’ view of IBM, Rockmann said it had become “a bit tainted by proceedings like this”.

Saunders suggested that if LzLabs had bought the mainframe under its own name then IBM, as a “highly litigious company would have learned LzLabs had its own mainframe and would want to know what use was made of it”.

Turning to an exploration of LzLabs executive emails, there was light relief when IBM was referred to in messages as a “dancing dinosaur”. That term was a reference to a book by a former IBM CEO, Rockmann said, presumably meaning Lou Gerstner’s Who Says Elephants Can’t Dance?, a 2002 memoir of the company’s corporate turnaround to become a nimbler player.

Yet another line of questioning concerned LzLabs’ keenness in an email exchange on Winsopia acquiring an IBM mainframe on the basis that IBM would not monitor usage. Rockmann stressed that this was because of cost efficiencies rather than a desire to conceal activities.

“That’s a complete misreading of what this email was saying [which] was that IBM would not be tipped off this was happening,” Saunders suggested. Rockmann said this was untrue and that, as a relatively small company, LzLabs was only keen to pursue best value.

Next, Saunders pursued LzLabs’ “code of conduct”. Rockmann said this began as a “belt and braces” approach designed to avoid inadvertent copyright infringements and ensure clean development processes with clear separation of roles. There was again a flurry of laughter in court when Saunders pressed Rockmann on potential meetings between LzLabs and Winsopia in shared office areas such as lavatories. Rockmann suggested that a server in a hole dug 330m in the ground would be probably safer but, more seriously, argued that LzLabs had gone to reasonable lengths to monitor compliance with its behavioural code, acting when it saw potential breaches. 

“A breach of the code of conduct does not equate to a breach of the ICA,” he added, referring to IBM’s customer agreement. Asked persistently about LzLabs monitoring procedures, Rockmann responded, “I have  no idea how you believe I can track every employee running around.”

Rockmann was also asked about an email from Moores, suggesting that “some things are failing, and failing badly” and that aspects of the “Chinese walls” policy has become “quite unworkable”. 

“We can book a trip to Farnborough for the celebration of the destruction of the Chinese walls,” Moores had written.

“That didn’t happen,” Rockmann deadpanned.

The case continues.

May 8, 2024: Day #6: IBM ethics questioned as trial turns to motivations in case against LzLabs

LzLabs lawyers question IBM’s intent when a special projects team investigated its Swiss rival

The latest instalment in the high-stakes trial examining whether a smaller rival broke contractual terms when it set out to develop an alternative source to run IBM mainframe workloads saw a marathon cross-examination of Big Blue special projects executive, Mark Anzani. Day six of the trial was dominated by cut and thrust between Anzani and LzLabs defence lawyer Roger Stewart KC, who carried on his work from yesterday by questioning Anzani for almost the whole of the day. Witness record-keeping habits, timings and suggestions that IBM was effectively trying to block an emerging competitor all came under scrutiny.

On Tuesday, Stewart probed “cloak and dagger” behaviour by Anzani and he began Wednesday’s proceedings on the same tack. Stewart suggested that Anzani’s habit of not maintaining notes or formal records suggested he did not conduct himself “in an honest way”, leaving “no record upon which to base your knowledge”. 

The absence of note-taking was “because of your experience of litigation … it can get you caught out with uncomfortable truths”, Stewart said.

As for timings, Stewart suggested that IBM and Anzani should have been familiar with LzLabs and its executive team at an early stage, as well as the fact that LzLabs is the parent of UK subsidiary, Winsopia. Winsopia’s acquisition of an IBM mainframe and subsequent examination of software code is central to the case.

However, Anzani said that matters only started to become serious when telecoms firm Swisscom went public as an LzLabs customer in 2019. While Swisscom undoubtedly was a lighthouse win and proof-point for the technology, LzLabs customers today include commercial vehicle maker MAN Truck & Bus, Canadian connectivity provider Rogers and Italian bank BPER Banca. Part of LzLabs’ defence is that IBM acted against it because it feared more such migrations to weaken its long-running mainframe cash cow.   

“The purpose of bringing this litigation was because clients were voting with their feet and IBM needed to stop LzLabs achieving business,” Stewart said. “Your litigation strategy has been to exert maximum pressure on LzLabs, Winsopia and [associated] individuals.”

Anzani said that the action was prompted by IBM’s own legal counsel and aligned with a long history of the company protecting its rights. But Stewart alleged that IBM’s actions were contrary to its own ethics code, employing “hordes of lawyers” to pursue a case counter to fair competition and innovation. LzLabs has said it spent over 300 million Swiss francs (about US$330m) developing its Software Defined Mainframe product that lets IBM mainframe users run workloads more affordably and flexibly.

On a warm day in London, one unfortunate, notable environmental aspect, consistent throughout the trial, continued. Enormous fans were in place to cool the well-attended court but the resulting noise made some remarks difficult to hear. 

The case continues.

May 7, 2024: Day #5: LzLabs rebuffs IBM claim it concealed Winsopia link and pushes its own ‘cloak and dagger’ claim

Details emerge of IBM counter-offensive versus mainframe software rival

Lawyers for LzLabs were kept busy in London’s Technology & Construction Court as a stream of witnesses were called to give evidence on day five of the IBM suit against mainframe rival, LzLabs. 

Much of that testimony concerned the relationship between Switzerland-headquartered LzLabs and its UK subsidiary Winsopia; IBM contends that the former attempted to conceal its relationship with the latter as part of its efforts to build a rival to IBM’s lucrative mainframe computing business. 

Farnborough, Hants.-based Winsopia is at the heart of the contract case as it acquired the mainframe. IBM says defendants went on to break terms of its licensing agreement amid research and development into providing an alternative way to run mainframe workloads.

Defence lawyer Roger Stewart KC began by questioning the relevance of the day’s first witness, Ian Lyon, a Big Blue sales veteran with over 30 years’ experience at the company. 

Stewart suggested that Lyon had only minimal involvement with Winsopia, was largely focused on hardware, was not focused on customer licence compliance monitoring, and had never checked the customer database that held details of Winsopia’s legal structure. 

Lyon agreed that, although he recalled being involved in a Winsopia hardware purchase, his “involvement in this one transaction was very minimal”. Stewart also noted that IBM had used a customer number for Winsopia with the corporate VAT number for LzLabs.

Next up was Paul Knight, a former IBM account manager until 2018. Stewart suggested that, as the customer account manager for Winsopia, Knight had visited Winsopia and must have seen employees wearing LzLabs logos. Knight denied this and also said he could not recall a 2013 email sent in his name offering Winsopia possible financing.

Stewart pursued a line of questioning around IBM familiarity with Winsopia, saying Knight must have accessed the relevant customer records database. However, Knight said that database was “absolutely hopeless” and had access rights controls attached. Pressed as to what Knight could and couldn’t see specifically, the witness responded, “I don’t know. I couldn’t see it.”

Emma Wright, IBM head of legal for UKI and Company Secretary, was called next. When asked who had sanctioned the current litigation action, echoes of the previous week’s evidence given by IBM’s Steven Wallin rang clear. Wright attempted to refuse to answer this question, citing privilege.  As with Wallin, Stewart took an aggressive position, raising his voice to say, “You’re a solicitor giving evidence under oath [and] refusing to answer a question is contempt of court”. 

At this, Wright relented, conceding that it was a senior figure within the company. She added that she did not know who had authorised the audit of Winsopia but conceded that Mark Anzani, a leader in IBM’s efforts to track LzLabs activities, had written relevant letters.

Matters took a literary turn as Anzani himself took the stand. Stewart, in a characteristically urbane comment, noted that Anzani had previously described his method of doing business as “cloak and dagger” with few notes or records retained. 

Stewart noted that the term derives from Chaucer: “With a knife hidden under a cloak. That’s you, isn’t it? You keep your knife hidden under your cloak of secrecy.”

Anzani said he was referring to keeping information “close to his chest” but conceded that not keeping formal records meant mistakes could be made.

Anzani said he had first heard of LzLabs in March 2016 and had almost immediately commenced an investigation into the company. That showed that US technology entrepreneur John Moores owned the company and led to the discovery of the link between LzLabs and Winsopia. 

The IBM investigation of LzLabs went under the name of Project Eiger and Anzani claimed that he viewed it “as an investigation to understand the technical capability of the ‘software mainframe’. 

“Some of that information became used on how to respond to client inquiries. I kept that away from the technical team working on the project.”

Stewart suggested that Anzani was also concerned with deterring customers from migrating from IBM to LzLabs and had gone “on attack, on occasion”. “On occasion,” Anzani responded.

The trial continues.

 May 2, 2024: Day #4: LzLabs lawyers turn up heat as first IBM witness gives evidence

Mainframe legal case sees doubts cast on Big Blue actions and motivations in suit against Swiss rival

On a day when the toasty setting of the Technology & Construction Court in London necessitated giant aircon units being drafted in between sessions, it was an IBM witness’s turn to feel the heat. 

In a long examination of IBM executive Steven Wallin, defence lawyer Roger Stewart KC suggested that Wallin had neither the seniority nor the knowledge to sign and send a warning letter to LzLabs’ UK subsidiary, Winsopia. That letter asked Winsopia to take rapid remedial action to comply with IBM’s ICA customer agreement covering appropriate usage of its software.

Under questioning, Wallin said he had no software auditing experience or in-depth knowledge of Project Eiger, pitched by LzLabs as IBM’s attempt to dissuade users from migrating mainframe workloads to its rival, until late in 2020. Wallin also had said he had only a “very high-level understanding” of terms of the ICA agreement.

Stewart said IBM had spent “an enormous amount of time since 2016” on Project Eiger but Wallin insisted he was unable to quantify the amount of time committed by the IT giant. Rather, he had been under the impression that IBM was merely investigating LzLabs’ activities, reviewing intellectual property, and attempting to understand the SDM product the Swiss company had created to allow customers to migrate critical workloads away from IBM. 

Quizzed about the Winsopia letter, Wallin denied that he had been working closely with Mark Anzani, an IBM executive who LzLabs suggests had led Eiger since April 2016. Eiger, LzLabs claims, was set up because LzLabs’s SDM had created a serious threat to its mainframe revenues stream. That stream is worth more than $3bn per quarter, according to a report cited by LzLabs.

“You knew full well Mr. Anzani had been conducting a campaign to keep customers away from LzLabs,” Stewart suggested. 

And, in a testy examination, he accused Wallin of “refusing to answer the question” as the latter frequently cited meetings held under “privilege”. 

Judge Mrs Justice O’Farrell intervened to remind Wallin he had to answer the questions. Stewart asked, “Has someone told you to hide behind privilege when asked difficult questions?” 

Later, he accused Wallin of “gabble” and asked why he was taking such long pauses before answering.

Stewart said that IBM had conducted a “sustained campaign to stop customers going to LzLabs” and suggested that Wallin had been a “stooge” for Anzani. Wallin denied working closely with Anzani, stating: “Anzani was a senior member of the leadership team. I was aware of him and interacted with him on a limited number of occasions.”

IBM, Stewart said, had pursued a policy of FUD, which he spelled out as “fear, upset and distress” to dissuade customers from adopting LzLabs technology. [More commonly, FUD is spelled out as “fear, uncertainty and doubt”. The phrase was popularised in technology circles by Gene Amdahl, whose company Amdahl marketed a mainframe competitor, to describe IBM sales tactics.] 

Stewart said IBM had also communicated to customers, warning of reduced security, performance and possible IP infringements. This was “plain FUD”, he said, and a threat that warned customers not to get on the “wrong side of IBM and its lawyers”.

Stewart said that IBM wanted to protect a revenue stream with extremely high margins of 80-90 per cent. Wallin responded by saying there have always been alternatives and healthy competition, as exemplified by the modern rise of cloud hyper-scalers.  

The case, scheduled to run for nine weeks, continues.

May 1, 2024: Day #3: LzLabs has its day in court and leans on EC Software Directive for support

Lawyers for IBM UK and its Swiss rival LzLabs debate customer agreement as arguments brew over contractual Ts and Cs

Lawyers for IBM UK and its much smaller mainframe computing rival LzLabs went deep into contractual terms and conditions on day three of their trial in London as the pair vied over rights pertaining to the IBM Customer Agreement (ICA) contract.

The case differs from many technology sector disputes in that it centres on whether LzLabs strayed from terms of the licensing contract. LzLabs lawyers argued over wording, for example suggesting that components such as macros or stubs (relatively tiny software elements) should not be treated as IBM programs. 

Users can observe, study and test code so long as those actions don’t infringe copyright, the defence argued, but there was no limitation on the extent of observing, studying and testing. So, in principle, the contract effectively allows the user to observe every internal working part of any IBM program, they claimed.

The defence contends that Lzlabs/Winsopia was allowed to examine code at a microscopic level, so long as it was not reproducing the “expression” of the code. “Expression” of the code is what is protected by copyright and not “ideas, practices, procedures”, it added.

Defences lawyers went on to argue that IBM’s case seemed to be that as soon as a customer includes a macro in the source code of its own application, it is restricted in its usage. “That’s commercial nonsense,” it said, contrary to the language of ICA, and would destroy the competitive ecosystem surrounding mainframe computing.

The claimants argue that Winsopia was systematically “looking under the bonnet”, using tools to exceed the parameters of observing, testing and studying because they dipped into the “internals” of IBM programs. 

The overarching defence is that code sent from Winsopia to LzLabs were not part of the ICA terms and so not a breach of contract under terms of the EC Software Directive. LzLabs also argued that it only pursued interoperability and, turning to allegations of disassembly of IBM code, it noted that the original four allegations had been reduced to three and said that this illustrates that suggestions of “systematic disassembly” are not supported by evidence. 

The case continues with expert witnesses for the prosecution due to take the stand.

April 30, 2024: Day #2: IBM, LzLabs lawyers fight fire with fire

On the second day of the closely watched trial, prosecution and defence exchanged barbs, claims and counter claims over what IBM knew and what LzLabs did

A fire alarm delayed proceedings at the Rolls Buildings in London on Tuesday amid more heated confrontations as IBM UK continued its claims against LzLabs for breach of its mainframe customer usage contract.

IBM lawyers said LzLabs-owned subsidiary Winsopia used a software tool as a “Swiss Army Knife” in order not to test how IBM mainframe applications work but rather to reverse-engineer code and thus support the development of LzLabs’ SDM (Software-Defined Mainframe) product. SDM lets customers run applications written for IBM’s host hardware on standard, affordable servers or in the cloud.

IBM also suggested that Winsopia’s “scrubbing” technique (that is, its cleansing of IBM code, data and attempts to remove errors) was flawed and inadequate. This process resulted in IBM materials being transferred to LzLabs and hence finding their way into the SDM product. Winsopia/LzLabs’ actions were “lousy and did a terrible job at removing IBM materials”, the prosecution alleged.

Also, building on claims yesterday that IBM did not know of the existence of Winsopia until 2020, prosecutors said LzLabs had deliberately concealed” the connection from IBM UK until 25 August of that year.

Lawyers for LzLabs responded, saying that argument stretched credulity as IBM had commissioned a taskforce to analyse the LzLabs threat as far back as 2016 and the connection between Winsopia and LzLabs was a matter of public documentation. An IBM representative had also asked for annual accounts for Winsopia and LzLabs as far back as 2017, it said.

“You have to assume that IBM were really stupid because you were trying to keep them from something they could find out really easily,” defence lawyers said.

The defence also said that the SDM program is composed of some 3.5 million lines of source code, all of which has been made available to the claimants. Not a single line was copied from IBM source, it insisted. Instead, it was more interested in providing customers with a choice of running mainframe software on a modern system.

It added that what it had created was analogous to a UK power socket where its SDM acts as an international adapter to access the energy source. It had used a highly experienced team of over 60 developers and had commenced work even before it had signed the ICA agreement with IBM, it claimed.

In a rare poetic moment, LzLabs’ lawyers quoted Thomas More, the lawyer, politician and satirical author of ‘Utopia’, who was executed as a traitor after falling out of favour with King Henry VIII. The defence doubtless believes that a victory would be Utopia for competition in the technology sector…

The case continues.

April 29, 2024: Day #1. Big Blue starts London trial with allegations of mainframe code misuse

Such was the state of the packed courtroom in the Rolls Building, London that the judge’s first spoken words were a request for further seating. The attendance levels reflected interest in, and the importance of, a crucial case between IBM UK and Switzerland-headquartered LzLabs with major potential implications for technology development and enterprise IT competition.  

After initial housekeeping discussions mostly concerned with timeliness of document submissions and requests for review, IBM UK prosecution lawyers went straight to their claim that LzLabs breached laws when its UK subsidiary Winsopia reverse-engineered mainframe code. That in turn, it said, enabled LzLabs to create its Software Defined Mainframe (SDM) product line that allows customers to run mainframe programs on commodity servers and in the cloud. 

That ability unlocks scope for enormous cost savings and provides customers with an off-ramp from Big Iron systems and onto modern hardware.

At its heart, the case brings together two polar views. IBM is painting the situation as “unlawful exploitation of technology” whereas LzLabs says that it has remained compliant with relevant laws including the EU Software Directive and UK law. It wants to foster a view that any judgement against it would be a blow to innovation and competition.

The first day of the trial saw IBM lawyers recount how Winsopia acquired a mainframe from a reseller and allege that the organisation went on to illegally break terms of the IBM Customer Agreement (ICA). “Winsopia repeatedly and extensively reverse-engineered mainframe software and provided the fruits [of its labour] to LzLabs,” lawyers said. 

The prosecution noted that several LzLabs figures also worked at Neon Enterprise Software, including serial entrepreneur John Moores, a defendant in the case. Neon and IBM also saw a legal clash in a case that ended in 2011 with Neon withdrawing its products from the market.

In opening remarks, IBM lawyers cited a statement released by a PR agency representing LZ Labs. They said IBM didn’t accept the depiction of it as “a monopoly charging exorbitant pricing” and adding that this theme was anyway not part of the trial. Instead, this was “simply noise by the defendant” and the case was “not Big Blue trying to pursue a competitor”.

In a notable statement, IBM also said it had only become aware of the existence of Winsopia in 2020, despite the unit having been set up several years previously. 

The case underlines the remarkable longevity of mainframe technology that has its roots in the middle of the 20^th century. All these years later, enterprise executives say mainframe modernisation remains a key concern with seven in 10 saying mainframes are central to their business technology strategies. According to whom? IBM itself in this 2023 report…